When I did my Computer Science degree the vast majority of courses were 50% final, 30% midterm - even programming exams were hand written, proctored by TAs in class or in the gymnasium - assignments/labs/projects were a small part of your grade but if you didn’t do them the likelihood you’d pass the term exams was pretty darn low.
In Spain, the whole university system was like that until like 15ish years ago. Exams were king, in most courses they were worth 80%-90%, and of course always in person.
Then we did a university reform, partly with the excuse of aligning with the rest of the EU within the Bologna process (and I say "excuse" because that's what it was, because the politicians introduced some things with that pretense that weren't like that in the rest of the EU at all, and it was perfectly possible to comply with Bologna without doing them) and partly to copy the US/UK ways. And one of the pillars of that reform was continuous assessment, and evaluating coursework.
As a consequence of this, first of all working class students were royally screwed. Because suddenly it wasn't OK to just organize yourself to prepare the exam, you had to attend lots of sessions to earn points, which put students who work at a disadvantage. And second, passing by cheating became possible, even before LLMs. People tend to forget that before everyone got access to ChatGPT, some people had access to experts (family members, or even paying someone to do the work).
Now that this kind of cheating has been democratized and everyone can do it instead of just the most privileged with access to experts or money to pay them, people act all outraged. Although pretty much nothing is being done, except for using snake oil detectors, or sometimes increasing difficulty of assignments to make them LLM-proof (with which you screw the students who actually want to learn without LLMs).
They spent years indoctrinating us (professors) in training courses on how the old exam-based ways were wrong (the "Napoleonic" model, they called it... none of them seems to entertain the thought that maybe if it had been working essentially unchanged since Napoleon it wasn't that bad, and you need solid reasons to change it beyond "this is old so let's change") and the new ways were the bee's knees. Like in the Milgram experiment, it's difficult for people to back down and acknowledge that they have been wrong, even when the solution is obvious.
I think the only defense of the new model is that it forces students to learn throughout the semester, rather than just before the exam. Which is easier and more effectively engages long term memory (like doing more rounds of spaced repetition).
I definitely could tell the difference, though most of the time I just studied full 4-7 days before the exam.
When I studied CS in Germany in the 2010s (also in the Bologna Bachelor/Master system) most courses had weekly graded assignments. But the assignments didn't count towards your grade, instead you needed to reach a certain total of points to be allowed to take the exam. The actual grade was entirely composed of midterm and final exam (pen and paper exams, no computers, no multiple choice).
It was easy to cheat on the assignments. Working on them in groups was common and sometimes encouraged. The only person you could really cheat was yourself (and a TA who had to grade one more exam)
In the UK they claimed that girls did worse on one-off exams and so the one-off exams structure favoured boys. When course had more graded coursework, girls did better.
So that was the justification used to switching to a less impactful final exam.
No idea how true that is.
We were also told learning a phonetic alphabet was better for young children learning to read than using the old ABC system.
As far as I have heard, that turned out to be based on one person's fantasy and zero evidence and has actually had negative impact on children learning to read.
The UK has a real problem with pseudoscientific nonsense invading the education system.
To my knowledge they still teach about audio/visual/kinetic learners and how you should structure the way you learn around which one you are. This has been debunked for decades.
> The UK has a real problem with pseudoscientific nonsense invading the education system.
Not just the UK, pedagogy/education is a very soft science, along with any other field that revolves around human behavior (psychology, sociology, etc...).
Using AIs in experiments and studies will be an improvement even if they do not accurately reflect human behavior, just because you don't need a harm review and you can repeat your experiments multiple times under different variables.
Yes, it does have some advantages. Apart from what you mention, another one is that it's not so consequential to e.g. sleep badly the night before an important exam. It's just that I find the disadvantages to be much greater than the advantages.
If seven days of study are sufficient to pass the class, why is so little material being taught in one semester? It sounds like the exams are far too easy.
The lack of any real innovation or major economic development in Spain would appear to be fairly strong evidence that the "Napoleonic" model wasn't actually working. Maybe this new system was even worse but I'm baffled as to why anyone would believe that changes weren't needed. Physician, heal thyself.
Spanish STEM graduates innovate just fine. They just do it abroad, where they get paid decent salaries if their work for others, or decent investment opportunities if they choose to be entrepreneurs :) Spain has lots of issues but I don't think education of the workforce has ever been one, neither before nor after the reform (it was definitely better before, but still, it's not half bad for now... Let's see what happens with the inaction with respect to LLMs).
In my public university in Spain, we always had the option to do a single final exam instead of the continuous assessment, although very few chose it. Generally the continuous assessment was less stressful, and the material stuck better, with room to digest it rather than just cramming for the exam and forgetting it right after. Generally the default expectation was that everyone was a full time student yes, but there were proper accommodations for those that weren't.
It is definitely a lot more work for the professors though, most of my family are teachers. It's a lot of assessments and it's very rare to have funding for TAs. Some think that the extra work with worthwhile for the sake of transmitting the knowledge more effectively, but not all of them do.
Frankly, you sound a bit bitter about it from the professor's perspective, and somewhat rationalizing why it is bad for the students. But students do generally appreciate it, and yes good students too, not just cheaters. I think both good and bad students end up learning more and hating the process less.
Your comments on Bologna do resonate though, it was very confusing when I continued to study in Germany and the Netherlands. The massive reforms were supposed to be for alignment with EU, but if anything it got more misaligned. They unified all 3 and 5 year degrees into 4 year degrees, but in most of EU all degrees are 3 years now, for instance.
Regarding the parent comment, indeed, my Computer Science degree was mostly hand-written exercises and exams, and it wasn't that long ago. The degree is about fundamentals, about understanding concepts and applying them, about the tools you need to learn anything in CS afterwards. You are expected to learn most of the practical skills for building software on your own, since they are ever-changing. And I have to say, that style of education has served me very well in my career.
PS: I was also surprised to learn that most of the undergrad exams in Germany, and some in NL, are oral. I can see how that might be a disadvantage to some, but writing is also a disadvantage to others. I quite liked it, less intense than a long written exam, and I think the professor can get a much clearer understanding of the student's grasp of the subject. But again, it's a ton of work for the professor, 20-30 mins per student one-on-one, giving them your full attention, adds up quickly.
Not sure when this was supposed to be the case, but for actual universities (not meant in a deragotary way, Germany has two types of higher eduction) in hard sciences, most classes are graded on a single written exam. Both in undergrad/bachelors and masters.
Unless things have drastically changed in the last five years...
This was in Freiburg University, which is among the top in Germany and top 250 globally. Computer Science bachelors and masters, around 8 years ago, most courses had a final oral exam unless many students (roughly >25) signed up to the class. Some classes like Information Retrieval or Machine Learning had >100 students so they had a written exam, but most others were smaller and oral: Data Engineering and Databases, Cryptography, Physics Simulations for Graphics, Formal Verification Methods, Bioinformatics, Planning AI, P2P Networking... I had a couple oral exams in VU Amsterdam (masters) too but fewer and not the final exam.
I know that an oral exam might seem less serious and rigorous, but I do think the professor can get a better grasp of how much the student actually understands the subject through an interactive interview.
My networking final in high school was probably my favorite test taking experience - there was a small written portion on eg subnets, but the bulk of our grade was setting up a physical network, testing it, and leaving the room. Our teacher sabotaged three parts of our network - could be hardware, router misconfiguration, etc. when we came in we had iirc 20-ish minutes to diagnose and fix it.
The best was when she barely unscrewed one of this big DIN connectors so at quick glance it looked fine, but wasn’t fully connected.
If it is mostly a ”show your work”/”show your reasoning” kind of grading where your width and depth of attempts are more important than success then it seems OK.
It also sounds vulnerable to bias, unless the teacher modified each project in the same ways. How did they ensure they gave each student an equally difficult challenge and didn't consciously or unconsciously make it easier and harder for different students?
Presumably they do the same thing for every student. I hope they aren’t unplugging a cable for one student while taking a sledgehammer to the next student’s router.
"For Miss Granger's ingenuity and resourcefulness, finding and patching a zero-day vulnerability in the router firmware... I award fifty points to Gryffindor."
When I did tertiary studies in programming there wasn't AI but we did our programming exams in pencil and paper. The "beneficial" prep we had and I had since high school was using punch cards. And 24h turnaround time for compiles. That really makes you think. And you learn how to desk check even thousand line programs. Intense focus, structuring for readability (to catch typos) and simplicity (catch logic errors) helped enormously. Was not unusual to change hundred lines of code and submit knowing that it wouldn't compile but will throw up the other errors I couldn't find. Our exams would give us 4-6 attempts for clean compile AND correct output. The only space where I experience same challenge now (40+ yrs later) is embedded code. Desktops and web stuff have LSPs and dynamic reloads and interpreted code (not a thing for me when learning) with instant feedback.
Lots of skills from those old days that have been lost/ignored in the pretence of productivity.
Yeah i really valued learning to code when I didn't have the internet available, if taught me patience and deep thinking, problem decomposition and organic (brain) execution
That is great for core principles. But languages and development environments have since assumed everyone has access to then internet. Meaning more "stuff" is the solution to problems (massive standard libraries or community created ones) rather that elegent language solutions.
The internet enabled all the complexity we have today. LLMs will have a similar effect, but instead of engineers actually having to understand the system (even in it's complexity) they will just be querying the oracle to build things or solve problems.
When the oracle can't help (or maybe refuses to) is when it gets interesting.
Similar to me, I learnt some html tags through a book which was sold at newsstand, once I was at my cousin's house using her computer without internet access, then I wrote a simple html page with the Win 95 cloud wallpaper as background image. My cousin was terrified how I did that!
24 hour turnaround for compiles, expected to change a hundred lines and it fail to compile, and you get 4-6 attempts, so up to 6 days to get a thousand line program working? Now you can compile and check a thousand line program in milliseconds. It seems so unarguably more productive, why do you call it a 'pretence' of being more productive?
> Now you can compile and check a thousand line program in milliseconds.
How do you define "productive?" Lines of code written per day? Bugs fixed per man hour? Fewest reported bugs per end user?
The fastest compiler in the world won't help you find all the runtime bugs that simply wouldn't have existed in the days of punch cards, when code was written with with more care and attentiveness since there wasn't a fast edit/compile/test development loop. YMMV, of course.
I personally dislike placing a heavy emphasis on exams. Assignments/projects have been consistently the most enjoyable and rewarding parts of the courses I've taken so far in university.
It's a shame that they are also way more susceptible to cheating with AI.
> It's a shame that they are also way more susceptible to cheating with AI.
They were more prone to cheating before AI, too.
Cheating has always existed at some level, but from talking to my couple of friends who teach undergrad level courses the attitudes of students toward cheating have been changing even before AI was everywhere. They would complain about cohorts coming through where cheating was obvious and rampant, combined with administrations who started going soft on cheating because they didn’t want to lose (paying) students.
AI has taken it further, with students justifying it not as cheating but as using tools at their disposal.
I was talking to my friend about this last week and he was frustrated that several of his students had submitted papers that had all the signs of ChatGPT output, so he asked them simple questions about their papers. Most of them “couldn’t remember” what they wrote about.
It’s strange to me because when I went to college getting caught cheating was a big problem that resulted in students getting put on probationary watch and being legitimately scared of the consequences. Now at many schools cheating is routine and students push the boundaries of what they can get their classes to accept because they have no fear of any punishment. YMMV depending on the institution
Those students will go on to "cheat" in job interviews. And the sad thing is that many of them will never face any consequences because the majority of jobs at large companies don't require any actual competence. The college degree allows them to check a box on job applications but the jobs don't actually depend on anything learned in college.
> because the majority of jobs at large companies don't require any actual competence.
An interesting side effect of the AI gold rush is that companies are starting to look critically at these do-nothing email jobs where someone forwards emails around and makes slides and Notion pages.
I’ve worked with many who occupied jobs that didn’t contribute much other than organizing text and sharing it around, but they got a pass because it looked helpful enough. Now it’s a lot harder to justify those positions when management realizes that having the not-really-competent person summarize communications and documents isn’t much better than having ChatGPT do it.
The dirty secret in HR is that performance in most regular jobs doesn't really depend on having learned anything in college. Companies only require a college degree to filter the candidate pool down to a manageable level. Line managers know they'll have to train entry level hires on almost everything.
IBM used to hire software developers based on aptitude test scores regardless of formal education, then put them through an extensive internal training program. It worked fine.
IBM (among other companies) used to have multi-month internal training programs based on their specific needs. And, in many cases, the expectation was that you'd be there for 25 years or whatever afterwards.
The problem with exams is that everyone has a bad experience with a poorly written one. Well-written exams will have questions that test students at different levels of understanding across the whole curriculum.
So a student who only understands the basics should be able to answer most of the easy questions and students who have a deeper understanding can answer the harder ones.
Well-written exams should feel pretty fair and leave students feeling like the result they got is proportional to the effort they put into studying the material (or at least how well they personally felt they understood the material).
Exams almost filtered me out of this industry before I even got started. I later went on to be a lead developer! I went to a rural high school with a poor math curriculum. I understood the concepts, but I was slow. When I got to undergrad, my first calc professor gave us a 60 question test with 50 minutes of allotted time. He told us if we couldn't do the problems fast enough we weren't cut out for the work and it would be better if we quit now. I've never felt more inadequate in my life. It's one of my only Ws.
> a 60 question test with 50 minutes of allotted time
Is this kind of test - many short questions - a standard thing for math in your country?
My university exams were pretty much all "2-question", in 90 minutes.
The first half was an essay where you have to reproduce a lesson from the curriculum, in your own words.
The second half was "the formulas" - you have to develop one or two formulas from first principles.
I once got an A- even though I got "the formulas" half very wrong. As the teacher explained later, I simply chose the coordinate system beginning at not the same place the textbook did. And this was supposed to be a bad teacher - he actually gave Ds to almost all of us (180 people). This was a makeup exam.
Yes. You can do it a set number of times, opting to either retake the course for a better grade or simply not take it again. Either way it shows on your transcript.
Exams are kind of like interviews, at some level they are always at least a bit artificial and sometimes very artificial. I don't really think they do a good job of proving understanding but we also don't really have anything better.
Every class. Teachers think: "We want a curve, not everyone is getting an A". This pushes the teacher to ask questions not covered in the material. If they stick to what is covered it is hard to get a decent curve. So, in the end, teachers just hand out grades based on IQ, and there isn't a point of grades.
Courses aren't supposed to be pure memorization. If you can only answer questions directly covered in the material but can't apply it to new situations, you should not get an A.
> Teachers think: "We want a curve, not everyone is getting an A". This pushes the teacher to ask questions not covered in the material. If they stick to what is covered it is hard to get a decent curve.
If you can't transfer the course's material onto new, unseen questions, then you might have memorised it, but you didn't understand it. Getting an A requires understanding, not rote memorisation.
You're forgetting the a large percentage of the student body simply does not care to judge the material fairly. Any difficult exam where they receive a low grade is perceived as unfair and complained about online. Professors that test fairly are reviewed harshly online. Universities that prioritize education over graduation receive poor reviews. Top universities specifically coddle and push their students making it nearly impossible to fail.
Also way more susceptible to cheating in traditional non-AI ways. And your mark ends up depending a lot on how much time you have to invest independent of how good you are at the course material.
Assignments and projects are great for learning, but suck for evaluation.
That is the traditional view, the view of those who want to improve their own knowledge and abilities, and presumably the view of those who would like to consider the degree to be a meaningful credential.
However I suspect that there are many who 1) are more concerned about the short term outcome, 2) consider the degree/diploma to be little more than a meal ticket or arbitrary gatekeeping without any connection to learning, 3) view the work as a pointless barrier to being handed said diploma, and/or 4) don't see the value of human learning in a world where jobs are done by AI and AI systems routinely outperform humans on complex tasks.
I'm curious about the objective competence of recent graduates compared to previous cohorts, not just in tech but in basic literacy and numeracy.
A lot of Gen Z are ferociously anti-AI, but for tribal and emotional reasons, not because of a nuanced understanding - which is ironic, because the nuanced reasons for being wary of AI are much stronger than the usual talking points about "stealing art".
Being tribal and emotional is going to make Gen Z easier to replace, because nuanced strategic insight is less common and more useful.
You say that but I was a Class of 2013, aka during the massive hiring boom of the teens. I tutored a friend of mine with a Ds get Degrees mentality who eventually graduated and now works an ass-in-seat job for Booz Allen or one of those types. I used to joke about it with another friend, that his diploma ought to include an asterisk and a half dozen other names for how much we ultimately did on his grades take homes. I’m pretty sure he makes about the same as me by now purely on tenure.
Personally, I dropped out despite a full ride+ becuase why would I put in work for a no name state school when I already has an FTE job as a developer out of high school anyway.
Turns out fraudulent action can still get the bag.
I agree with your premise about why accurate evaluation matters, but your post comes across as pretty bitter. Unless you’re at the job with him, you really don’t know that it’s a “I just need to show up” job he has at Booz Allen. Perhaps he has other great traits like a high social or emotional intelligence that make him good at his job beyond whatever was being evaluated on those projects you helped him with.
I think you're missing the point. The majority of jobs at companies like Booz Allen are sort of like Kabuki theater and don't require any technical competence. The main responsibilities are to show up on time and present a certain image to customers.
Part of the purpose for evaluation is to provide feedback. I'm not going to claim that the form of feedback is great, but it does offer motivation to improve.
The other thing that feedback feeds into is credentials. I realize that some people are dismissive of this aspect of the degree, but it is important to pursue further studies or secure a job. While you can argue that these people are only cheating themselves, and some of them are cheating themselves, a great many will continue to cheat as they advance in academia or the workforce. In other words, they are cheating others out of opportunities.
You can imagine a world where the Corporate or State AI handles education, tailors it to individual student levels and talents, and assigns work based on its own direct experience of a lifetime of interaction.
You can also imagine that in that world where most humans would be redundant - unless the AI was optimising for human-to-human jobs and for evidence of unusual insight and creativity, not managing bullshit work for corporate profit.
The problem is that Universities are no longer institutions where people who want to can go to educate themselves but have instead been turned into the first part of your job interview. In this world, Universities are judged on how well they evaluate their students.
I went to college as a MechE so unsure if compsci was different. But overall, all the “fun” projects were labs. We have three semesters of hell and all 3 semesters had 2-3 labs, and we write 20 pages or so for EACH lab a week (usually a team of 2-3).
I also think assignments and projects work best, but only for the percentage of students that are self-motivated, curious, interested, etc.
Unfortunately a lot aren't, they feel like they have to be there or these courses are the only path for them to get a good job. And unfortunately they end up in the workforce, too. You'll often see teams with one good developer and a lot of hangers-on.
Then I suppose we can go back to having computer labs that can only access white listed domains and other study materials. Students code there to ensure no cheating.
The labs I was in weren't connected to the Internet at all, only a local intranet. Though, they were all running pre-oracle solaris if memory serves, so I'm probably dating myself a bit.
That was common when I studied computer science in the nineties as well. Hand written exams mostly.
Writing papers is a useful skill to have. And many students aren't very good at that. I taught some classes during my Ph. D. and supervised some students with their master thesis and PhD thesis work. Many students get their degrees without that really getting addressed. At least Computer science degrees in the Netherlands just spend very little time on writing skills. You get students with high school levels of English and Dutch and that's it.
I learned to write properly only when I started my Ph. D. My supervisor made me do it right before he allowed me to submit papers for publication.
AI might actually be good for education long term. It will result in a more personalized approach, which I think is good. There are plenty of ways to test students that are more engaging and interesting for both teachers and students than some of the old ways. You can't fake knowledge when you do a verbal test. Or test people with a good old written exam.
And of course for teachers, you can automate a lot of the verification work. This can be a lot of work.
Yeah none of the problems with AI in education are new; some schools (or news articles) are just panicking because they gave their students laptops (and/or made them a mandatory part) and now the genie is out of the bottle.
But there were already heaps of problems with tech in education before AI.
My CS projects were often pretty free-form so in theory I could've just used AI - today, anyway. But a big part of the grade was a face to face interview where you actually had to talk about the code you wrote. Anyone lifting along with other people who didn't actually do any work would fall through then.
Which strikes me as a terrible way to teach and test programming skills. If you're teaching to program without so much as syntax highlighting, you're not preparing your students for anything that even remotely resembles the industry they aspire to work in.
Honestly, these days universities should probably find a way to incorporate AI into their teaching, rather than fight it. Anything else is betting that AI will not stick around, which strikes me as a hopelessly naïve bet. Especially for software development.
I don't pretend to have all the answers, I don't know how to teach systems thinking in a appropriate way either. But I'm pretty sure typewriters isn't it, unless your students are hoping to get hired by Ada Lovelace, it's just not going to be relevant.
Education isn't any churning out a cog fitted to a bigger gear. A college education should not be preparing you for intellij or familiarizing you with va code.
It should be teaching foundational skills with the discipline. And I just don't understand how you look at the amazing things we've accomplished since ww2 and say "the education system that taught this fucking sucks"
What I don't understand is how you can look at all the progress we've made since ww2 and decide we don't have to update our education to keep it relevant, and argue in favour of teaching tech that's been obsolete for 80 years.
Students today aren't studying to repeat the progress made since the forties, that's been done by their grandparents, they're looking to drive the next 40 years of innovation, and if that involves typewriters I'll eat my hat
That way of teaching got us to the moon, created transistors, produced the internet, smartphones, quantum computers, the very AI that everyone is talking about, vaccines, sent probes into space, cured diseases, fed millions, jumbo jets, basically every single thing that society has come to depend on.
IMO, it's a bit disingenuous in 2026 to argue that technical college degrees exist for the sake of pure education and that we're best suited to continue pretending as such.
The vast majority of students and employers treat them as vocational certificates in practice, and the profession would almost certainly benefit from adapting curricula to more closely match that reality.
Foundational concepts are still necessary, but I don't buy the argument that we should continue teaching like it's 1946.
Syntax seems like a stupid thing to test in university level courses. That's trade school stuff. And I don't mean that as a criticism of trade schools, they just have a different focus.
Syntax is not the focus of your testing, but it’s often a pre-requisite to be clearly and accurately speaking the same language. Think not of taking off points for missing a semicolon but instead understanding the difference between the syntax for a method call and a property access. The different syntax conveys different meaning and so we should expect some basic level of accuracy to the language in question. At least that’s how I see it.
Marking at scale is hard to maintain that consistency though. It’s not whether the exam writer sees it that way, it’s whether the markers understand intent and objective over pedantic nuance
Knowledge is built on foundations. Knowing syntax in one language is necessary to be able to do anything practical, which interacts with theory. You build valuable schema of the world by iterative theory and practice.
College exam tests C++ syntax and people frame it as a good thing. Interesting. It sounds almost indefensible for me. If it were Scheme perhaps it would be okay, but C++ and Java...
We are at a point in AI where colleges teaching computer science are having students hand write code. It is what it is. This is about academic honesty^H^H^H^H^H^H^H^H integrity. In order to measure academic success we have to measure the person learning, not the machinery.
When I was in university we had the "Honor System"—no proctors, but you had to sign a statement on every exam and assignment that you did not cheat. And if you did cheat, one of your fellow students could report you to the Honor Board. Basically using the prisoners' dilemma to enforce honesty. And the Honor Board were always threatening to bring proctors back if cheating continued.
But yeah, everything was hand-written. On sheets of paper with pencil. I even had to write x86 assembly out by hand for my CPU architecture class. Of course, laptops were available back then but not cellphones and certainly not LLMs, so cheating by electronic means probably presents a stickier wicket now than it did back then.
The prisoner's dilemma requires that both people are facing the same consequences and are promised leniency if they give up the other person. I'm struggling to figure out how this conforms to a person giving up a cheater to the honor board. What am I missing?
I’m guessing we’re similar vintage. My CS classes were like this as well.
The only exception is that when I got into grad level classes we did have some big programming projects. But most of that programming happened on sparc stations, and it was actually just easier and more productive to sit at the machine in person with its nice big (at the time) display with all the other folks doing programming projects. Those machines had the standard dev toolchains provisioned that weren’t easy (at the time) to do on a dorm room Mac or windows computer.
I really think a lot of the ways we can reduce reliance on AI for thinking is to just set up systems where it’s not an inviting or rewarding option.
Same. Coding by hand was really, really hard at first, I remember. Then one day it just clicked and it was like I could think so clearly in my head, without making mistakes, and write code that would have compiled by hand during an exam. We would just practice and work for it.
I don't want to be polemic, but I really miss those days.
Yeah writing out code by hand made me slow down and think more.
One thing I recall is that the grading policy made it very clear that minor syntax issues were inconsequential in handwritten answers. And more advanced classes only wanted pseudocode. Which are exactly the right priorities.
Not only that, the tabletification of education has actually been shown (via standardized testing) to make our kids dumber. This is the first time a new generation has scored worse than their parents. Technology has its place, we need to pick and choose where.
Half related: reminds me of my physics teacher's test of how observant we were. The extra credit question on the test was "what is your teacher's favorite color?", which she had so far given no indication of. But while watching us she was walking all over the room in every possible direction, because the answer was on a piece of paper taped to her back.
My dad had a class in medical school where the professor dipped his finger in a beaker, talked about mechanisms for sugar in diabetic patients’ urine, and then proceeded to stick his finger in his mouth. He had all of the students do the same, who noted that the sugar was apparent.
He concluded the class by talking about the importance of observing patients, and pointed out that he had tasted a different finger than the one he had put in the beaker.
You mean an AI proof grading system. Grades have very little to do with actual learning and are there solely for signalling. It would be better if Universities shifted their focus to learning and eliminated grades/exams all together. They should seek to stop trying to be the gatekeepers for white collar work and instead focus on learning and research.
My local college used to have a test center you would physically go into. Now they no longer let teachers send their students there (idk if its just for programming or what) for whatever reason. I know they've wanted to cut down on wasted paper for years.
If my college is doing this, I cannot imagine how many others are also impeding on their entire goal: education.
Seems like anybody could just study 6 hours a day for the last month before the final, last 2 weeks before the mid term and use AI for everything else.
Some people spend ages writing a cheat sheet with the intent of cheating on the test but realise that because they wrote it down and / or tried to summarise it they actually learned the materials.
I used to make my classes 60-80% project work, 40-80% quizzes all online.
I now do 50% project work, 50% in person quizzes, pencil on paper on page of notes.
I'm increasingly going to paper-driven workflows as well, becoming an expert with the department printer, printing computer science papers for students to read and annotate in class, etc.
Ironically, the traditional bureaucratic lag in university might actually help: we still have a lot of infrastructure for this sort of thing, and university degrees may actually signal competence-beyond-ai-prompting in the future.
I always preferred the "you get some grades along the way to gauge your progress but the lion's share of the weight went to the proctored exams" method unless the lion's share of the normal work was also proctored anyways (at which point it doesn't really matter how it's done).
The reason was less for myself and more because anything group related suddenly shot up in quality when the other individual work classmates were graded on couldn't be fudged.
The things I don’t like about putting too much weight in the exams are:
* It’s sort of unnecessarily high stakes for the students; a couple hours to determine your grade for many hours of studying.
* It’s pretty artificial in general; in “real life” you have the ability to go around online and look for sources. This puts a pretty low ceiling on the level of complexity you can actually throw at them.
Exams happen all the time in real life. Or rather, situations where you can't just look up fundamental knowledge. Job interviews, presentations, even mundane work tasks - all these require you to know the basics quickly "The basics" are relative, of course, but I often point out to my students: "you don't care if your doctor needs to look up the specific interactions of your various meds. You do care if you see them googling 'what is an appendix'." Proctored, in-person exams are the only reliable mechanism we have for ascertaining if a specific individual has mastered key fundamentals and can answer relevant questions about them in a relatively timely fashion. Everything else is details and thresholds - how fast do you need to be able to recall, how deep, what details are fundamental. From there, I think it's fine to hate poorly made exams, and it's a given that many folks making exams have no idea what they're doing (or don't have the resources to do it right). But the premise of an exam is not completely divorced from reality.
I think many of us would agree that job interviews (in tech at least) are horribly broken, because they don't do a good job of testing candidates' ability to do the actual work they'll be doing day-to-day. So saying exams are like job interviews is not a positive for exams. And, for most people, the ideal is to find a job and stick with it for years, so it's not like job interviews are common, everyday occurrences.
For presentations, usually you spend a lot of time preparing for them (similar to exams), building a slide deck or pages of notes that you refer to while giving the talk (not similar to exams). Sure, you do have to be able to think on your feet, but I don't think the comparison to a sit-down exam is all that apt.
For mundane work tasks, you have the internet and whatever reference materials you want (including LLMs, these days); this sort of thing is so different from a sit-down exam that it's almost comical that you'd try to equate the two.
I'm not saying I know of a better way to evaluate learning than proctored, in-person exams, but suggesting that sort of situation is particularly relevant to real life... no, no way.
Having been both a data analyst and software engineer I agree. The data analyst one? Here is 50K of Excel rows with all kinds of weirdness in it, you're data analyst right? You have 4 hours to analyze this data. Go!
The software engineer one: here is a takehome assignment. One week later: finished!
To be fair, they both represented pretty well what work I'm going to do. The data analyst didn't show that well how much I'd also be data engineering, but whatever, I was a SWE before having a DA stint. Back to SWE again though.
>you don't care if your doctor needs to look up the specific interactions of your various meds. You do care if you see them googling 'what is an appendix'.
Sounds like you're saying that it's acceptable to be a little foggy about the limits of your knowledge, as long as you remember the core foundations. For a first year medicine student, the edge of their knowledge will include things that are the core foundation of a practicing doctor. Why should such a student be tested as if he already had several years of familiarity with the subject when this is all relatively new material to him?
I don’t completely disagree with this and I do appreciate the well thought out response. I should say that most of my recent relevant experience was as a grad student teaching assistant so I’m neither just somebody grouching that I didn’t like taking exams, nor somebody with deep experience setting the things up. Some respectful quibbles, then, since I gather you have more direct experience than me.
> Proctored, in-person exams are the only reliable mechanism we have for ascertaining if a specific individual has mastered key fundamentals and can answer relevant questions about them in a relatively timely fashion. Everything else is details and thresholds - how fast do you need to be able to recall, how deep, what details are fundamental.
I don’t think this is how people actually engage with exams. I had a lot of folks in office hours who treat the exam as the ceiling of their competence, rather than the floor, and do things like cram or try to figure out exactly what topics will be on the test to study just those. If the goal is to establish a 100% solid foundation for things you have to know to be a professional (which I think is a great goal), I prefer something like Mastery Learning to the conventional exam process. (Maybe we could call Mastery Learning conventional exams a different set of thresholds, unusual thresholds if we want to look at it that way).
> From there, I think it's fine to hate poorly made exams, and it's a given that many folks making exams have no idea what they're doing (or don't have the resources to do it right). But the premise of an exam is not completely divorced from reality.
I worked with some professors who I thought gave good exams, some who gave less good ones, so I don’t think the premise is completely divorced from reality. But it seems more like something the good instructors overcame, rather than a construct that is really helpful.
I think it's all about speed. In "real life" everything can be looked up, but exam optimizes to not even having to look it up. Then any research becomes much faster.
Whether it's good or bad I don't know, I think US higher education focuses too much on ability to produce huge amounts of mediocre work, but that's the idea behind exams.
One of the reasons I've always encouraged software people to learn to touch type has nothing to do with typing speed - it's about reducing/eliminating the cognitive load of typing, you want to be thinking in expressions (sentences) not letters. (The increase in effectiveness comes from not getting distracted by the mechanics of typing...)
In real life you need to know the options and their trade-offs to solve a given problem. You don't need to know all the techniques perfectly, but you do need to be able to characterize them and compare them, from rote memory.
I agree, I think many people who rail against exams underestimate how important memory is to more complicated skills. How can you debug a complex application if you have to keep looking up every operator and keyword in the language you're using? It'd be like trying to interpret poetry in a foreign language but you have to look up every single noun. I'm not saying people can't do it, but it's tedious, slow, and you probably wouldn't think of them as a "professional worth paying for their service". Some amount of memorization is key.
It still doesn't feel to me that those things are similar. A sit-down exam is a time-limited, high-pressure situation where you're expected to demonstrate proficiency in the things you've learned over the past several months. Sure, much of that learning builds on stuff you've learned previously, but the focus is on the prior semester (or half-semeter, for mid-terms).
When I sit down to debug a complex application, I'm drawing my prior 25+ years of experience. While I certainly would rather fix the problem faster rather than slower, I don't have a time limit, and usually taking my time (or even leaving the problem alone for hours or days) can be more effective than trying to work quickly and get everything done immediate.
The last time I sat for an exam was in 2003, and I honestly have not experienced anything in life since then that feels like that. Even job interviews have not felt similar enough to me to evoke that same feeling. (Frankly, I've enjoyed most job interviews; I don't think I've ever enjoyed an exam.) That's just my experience, of course, but I don't feel like an outlier.
> It’s pretty artificial in general; in “real life” you have the ability to go around online and look for sources.
Sort of. In real life, you are expected to have immediate knowledge of your field and (in some environments) be able to perform under pressure. I'm not going to pretend the curriculum is a perfect match for what people should know, but it does provide a common baseline to be able to have a common point of reference when communicating with colleagues. I would suggest the most artificial thing about exams is the format.
> It’s sort of unnecessarily high stakes for the students; a couple hours to determine your grade for many hours of studying.
I don't like dismissing the ordeal of people who face test anxiety, but tests are not really high stakes. There is a potential that a person will have to repeat a course if it is a requirement for their degree. At least at the institutions I attended, the grade distribution across exams and assignments, combined with a late drop date, meant that failing a course was only an option if you choose it to be. A student may be forced to face some realities about their dedication/priorities, work habits, time management, interests, abilities, etc.. It may force a student to make some hard decisions about where they want their life to lead, but it does not bar them from success in life. And those are the worse case scenarios. A more typical scenario is that you end up with a lower GPA.
>There is a potential that a person will have to repeat a course if it is a requirement for their degree.
How is that not high stakes? The result of several months worth of effort hinges on what they do during a 2-3-hour window. If you had to build something and the last step involved a procedure that could potentially tear down everything you made, you would try to find a way to redesign it so that didn't happen, or to limit the scope of the damage.
If you want an easy, stress free job, sure, nothing is high stakes. You will never be invited to meetings with executives, and you can google everything all the time.
If you, on the other hand, attend meetings, you will need both deep knowledge (requires many hours of styding) and fast thinking, and the questions will make you realize you know little about many things. You need to have general knowledge of many things, not just whatever you are building at the moment. If you are successful on these meetings, your salary can and will grow.
Also, some meetings will make you realize that college and/or university are life in easy mode. Very few things you consider hard in college will be work-level hard.
This is where the alternative of a course with the other (still monitored for graded activities) option comes in. The downside of that tends to force in person synchronous rather than custom scheduling of regular tests.
The point is more about whether the graded work is actively reviewed than which individual choice is ideal or not though. Whether it's electronic or written, remote or in person, weighted towards exams vs continuous are all orthogonal debates to the problem of cheating/falsely claiming work.
I had attended a few courses over a decade ago and just completed a degree recently. The methods of cheating have changed, but not because of pencils vs keyboards.
High stakes artificial exams can help prepare you for artificial stakes at job interviews where you need to crank out a working solution in 30 mins with jet lag and someone looking over your shoulder
That's true. They do better-prepare an applicant for a job that filters on a person's ability to accomplish arbitrary things in a vacuum that is completely disconnected from the real world.
That's probably a good thing to filter on for, say, the navigation role on all kinds of crafts (from land to sea to space). There are naval roles where navigating with a sextant and memory is an important skill to have, and to test for.
But that operating-in-a-vacuum skill doesn't relate well to roles that don't need to exist in a vacuum. In most of the jobs in the real world, we get to use tools -- and when the tools go out to lunch, we don't revert to the Old Ways.
When an accountant's computer dies, they don't transition back to written arithmetic and paper ledgers. Instead, someone who fixes computers gets it going again, and they get back to work as soon as that's done.
Obviously they're both supposed to be proxy measures, not realistic scenarios. I was mostly joking before but I do think exams provide a pretty good proxy for ability in the subject if the teacher is decent. Interviews not so much unless the applicant is similarly prepared with foreknowledge of what they will be tested on and had some time to prepare and given recent practice.
> It’s pretty artificial in general; in “real life” you have the ability to go around online and look for sources.
I dunno how you work, but I'd be getting raised eyebrows from people watching my hit google for any question required of my role.
I mean, we're not talking about using calculators here, and we're not talking about vocational training (How do I do $FOO, in docker? In K8s? How do I write a GH runner? Basically any question that involves some million-dollar company's product).
We're talking about college stuff; you absolutely should not be allowed to look up linked lists for the first time during an exam, copy the implementation from wikipedia, port it to your language and move on.
In the real world, we want people who mostly know what to do. The real world is time-constrained (you could spend 2 hours learning to do what they thought you could do based on your diploma, but they'd be pissed to find out that you need to look up everything because that's how you coasted through college).
Exam situations are more like the real world than take-home assignments: High-stakes, high-pressure, timeboxed.
If your real world does not have high-stakes, high-pressure, timeboxed tasks, then you really haven't had much contact outside of your bubble.
So at 50%, someone who uses AI to get 100% of the homework grade will earn a D (sometimes passing) if they can get at least a 20% on your quizzes, and a C (always passing) if they get at least a 40%. Did you make your exam so difficult that students who truly didn't learn the material earn less than 20-40%? Because if it was, say, multiple choice questions with four possible answers, then you can expect them to earn at least 25% just by chance.
While that answers their direct question, they do bring up a good point -- how often are you handing out less than 25% scores on exams? Id imagine any professor to do that to get some severe criticism that would make even a cheater pretty livid
What's interesting is that as I understand, folks are using things like Google Docs for papers, and that it's (apparently) straight forward to do analysis on a Google Doc to see, well, the life of the document. How it was typed in, how fast, what was pasted and cut back out.
My understanding is that the Google Doc is not a word processing document, it's an event recording of a word processor. So, in theory, you could just "play back" watching the document being typed in and built to "see" how it was done.
I only mention this because given the AIs, I'm sure even with a typewriter, it's more efficient to have the AI do the work, and then just "type it in" to the typewriter, which kind of invalidates the entire purpose of it in the first place.
The typing in part is inevitable. May as well have a "perfect first draft" to type it in from in the first place.
And we won't mention the old retro interfaces that let you plug in a IBM Selectric as a printer for your computer. (My favorite was a bunch of solenoids mounted above the keys -- functional, but, boy, what a hack.)
TaaS -- Typing as a service. Send us your Markdown file and receive a typed up, double spaced copy via express shipping the next day!
This would take about 1 day for some student to realize you can instruct one of the LLMs to operate the computer screen for you and have it type and fake edit a document for you. The tip would spread among the cheaters and the metric would become harder to judge by itself.
That's certainly one way to abstractly automate a task: Just pay someone else to do it. (This is a concept that regular people employ every day in the real world.)
Another way to automate this particular task is that some typewriters have (serial/parallel) ports to connect to a computer. It's not a daunting task at all for a student who is skilled in the art of using the bot to have one of these typewrites be the output target.
Even Microsoft Word stores revision history inside .docx files, and that’s been used to expose plagiarism. I heard about one case where a student took an existing paper (I believe from a previous year/student) and pasted it into Word. They then edited it just enough to make it look different.
However, they didn’t remove the embedded revision history in the .docx file they submitted, so that went about as well as you can expect.
Are you sure about that? I could easily see this happen with a web document link, but for a docx file the change tracking is off by default and pretty obtrusive. Basic metadata would be fine, formatting might be quirky but that's not exactly a smoking gun...
It’s been a while since I heard about it, but IIRC the professor was a stickler for a very specific paper format, so they would distribute a .docx template file with Track Changes already enabled and require students to write their papers using that template.
I also think that when track changes was first introduced in earlier versions of MS Word, there wasn’t as much concern about privacy/telemetry as there is now, so it wasn’t made as prominently obvious.
I'd be surprised if copy/paste carries the revision history, though. Wouldn't they have had to start with the original document (from the other student) and make their edits directly, and then submit that file?
I think they mean the revision history in the submitted .docx file showed the copied paper being inserted (as a single change), and then various changes to attempt to hide the plagiarism.
Hmm, I have some old daisy-wheel printers in the closet that I've been meaning to strip down for stepper motors, maybe I should refurb them instead :-)
You should look up the history of the Loebner Prize [1]. There’s a shocking amount of technological development in some chatbots that went toward simulating mistakes and typing patterns to make them seem more human-like.
In some of the later Loebner competitions, when text was transmitted to the human character by character, the bot would even simulate typos followed by backspacing on screen to make it look more realistic.
Yeah I definitely think LLMs contributed to its demise. To be honest, nobody in academic AI circles took it very seriously, because it kind of devolved into a contest over who could create the most convincing illusion of intelligence.
Participants spent more time polishing up the natural language parsing aspects in conjunction with pre‑programming elaborate backstories for their chatbot's bios among other psychological tricks. In the end, the whole competition was more impressive as a social engineering exercise, since the real goal kinda became: how can I trick people into thinking my chatbot is a human?
But reading through some of the previous competition chatbot transcripts still makes for fascinating reading.
Hi. I got a notification for this message in my email, so thought I'd reply. My name is Steve Worswick and I won the Loebner Prize a record 5 times with my entry Mitsuku between 2013 to the last contest in 2019. The Loebner Prize didn't take part in 2020 due to Covid restrictions and never started back up again. This along with the death of Hugh Loebner and the resulting end of his sponsorship is why the contest ended. LLMs started to become more popular in 2020 but this was pure coincidence.
I'm old enough to remember a similar controversy over whether to allow calculators in math classes. While most schools were banning them to force kids to learn how to do math without them, my school went the other way. They mandated that every student had one and then changed the assignments and tests to account for it. Gone were questions that had whole number answers that could be computed in our heads. Instead, answers were complex and the only way to know whether you'd done the question correctly was to be sure of your method. They even allowed us to write programs in TI-BASIC that we could use on tests, the only limitation was that we were not allowed to share programs with other students. I discovered that rather than trying to cram for exams, I could just write a program that would solve each class of problem we were likely to see on the exam, and by essentially teaching my calculator to pass the test, I also taught myself. It was a vastly better way for me to study. It also led to my decision to major in comp sci and my career in software. I'm forever grateful to those teachers for choosing to see the latest technology as a multiplier of student potential rather than a way students could cheat to avoid learning.
So I can't help but wonder whether schools are going about this all wrong. Rather than banning the use of AI and trying to catch students who are cheating, why aren't they creating schoolwork that requires AI? These tools are not going to cease to exist. The students they are preparing are going to live and work in a world where they exist. To my mind, you best prepare students by teaching them how to use the tools most effectively, not by teaching them how to work without the tools. Students should be learning how to prompt AI without hinting it towards a specific answer. They should be learning how to double check the answers AI gives them to ferret out hallucinations. They should be learning how to produce work that is a hundred times more complex than what us older folks had to do in school. We should be graduating students who are so much more capable than any generation before them. I think we're doing them a disservice by trying to give them the same education that was given to those from previous generations. The world they will inhabit has changed radically from the one we entered into following school.
> So I can't help but wonder whether schools are going about this all wrong. Rather than banning the use of AI and trying to catch students who are cheating, why aren't they creating schoolwork that requires AI?
Because using AI is the complete opposite of "I learned programming just to make tests easier".
By learning how to program solver, you not only learned how to program but also learned the method well enough to write it.
By pawning it off to AI to solve, you have learned nothing, not even how to prompt correctly as test questions are usually formulated well enough that AI doesn't need prompt massaging to get it.
You can use AI to get some knowledge about the problem (assuming you won't hit hallucination) but that's not what will happen when you use it for test.
And if you DO want to teach students how to use AI effectively, you can just have an AI class...
>By pawning it off to AI to solve, you have learned nothing, not even how to prompt correctly as test questions are usually formulated well enough that AI doesn't need prompt massaging to get it.
If you got AI to produce a working solution, you solved the problem. In the real world nobody who's paying you cares about the method as long as you deliver results. Students taught to solve easy problems by themselves will be at a big disadvantage in the workforce compared to students taught to solve hard problems using AI.
The part you're missing is that the evaluator already knows the answer. They're not looking that you can arrive at the correct answer, but that you know how to arrive at the correct answer. If "arriving at the correct answer" just means retrieving data from a Baysian database using a Markov chain you have only demonstrated you provide no value in the chain and should indeed get a mark of zero or get recycled.
>The part you're missing is that the evaluator already knows the answer. They're not looking that you can arrive at the correct answer, but that you know how to arrive at the correct answer.
The university evaluator is not the one paying you, the one paying you is your boss or customer. It doesn't matter how highly your university professor thinks of you, if you can't solve difficult problems as fast because your university never taught you to solve hard problems with AI, you're going to be at a competitive disadvantage in the workforce when you graduate.
I don't think the entire purpose of schools is to teach you how to answer some specific set of questions; they want to improve your knowledge and skills in various domains, and the questions are merely a roundabout way to assess that. If you can answer the questions but don't know the knowledge, you're missing the most important part.
For most students the "purpose" of studying computer science at university is to get a better job and make more money. And for the people for whom this isn't the case, they're generally smart and motivated enough to learn the extra details they're interested in by themselves.
There's also no reason to learn to read and write! First graders could just point their phone at some text and have it read to them, or dictate to their phone to achieve the reverse. Why learn to swim, walk, run? Machines can do that for you too!
For now there's plenty of people who are significantly more capable than AI models. Someone who fully outscources to machines will never join that club.
You have to evaluate students on their own skills before you continue their education, because at some point AI models won't be able to help them. Anyone can use some LLM to pass the first few months of undergraduate engineering disciplines, but if you got through that and haven't learned a thing, you're completely fucked. Worse, you won't even notice the point at which AI starts to fail until you get your test results.
Once the above is not true anymore, education is pointless anyways. However for now AI can at best replace the worst performers and only in some areas.
>You have to evaluate students on their own skills before you continue their education, because at some point AI models won't be able to help them.
If at some point AI models won't be able to help them, then give them assignments that reach the point where AI alone isn't enough, so they'll only be able to solve them if they learn whatever is necessary. This is what's meant by "making assignments harder". Students who learn to solve harder problems with AI will be more competitive in the workforce than students who only learn to solve easier problems by themselves. Because AI already allows people to solve harder problems than they could unassisted, but it's a skill that needs to be learned.
As an example, with AI, it'd be a reasonable assignment to ask students to write a working C compiler from scratch. Without AI that'd be completely beyond the reach of the vast majority of students.
That's great for autodidacts, but most students will be stumped by a complicated problem if you don't slowly walk them up an incline first.
Also what do you think is an appropriate assignment for first graders where "AI is not enough"? Are we supposed to give them problems meant for engineering majors?
The things you are saying at best apply to a few select areas of education and you are hyperfocusing on them. What you are neglecting is that a lot of education focuses on teaching tool use: reading and writing is a tool, CAD software is a tool, AI is a tool, even language is a tool. For many people the best way to learn to use tools is being taught by another human being. That human being has to evaluate their progress somehow. If a first grader uses their phone to have text read to them, this tells me very little, except maybe that they can at least understand spoken language to a degree.
Using LLMs effectively, especially without essentially becoming the LLMs meat-puppet, requires a set of skills many 10th graders still struggle with. Skills like putting what you mean into words, extracting meaning from text, and thinking critically about the information you are fed.
Finally there's the matter of philosophy, ethics, and politics, which also happen to be on the curriculum in some places. Are you going to let a LLM argue for you? If you have never learned to evaluate your own beliefs and turn them into something coherent that you can communicate to others, and instead let the LLM argue on your behalf, then congratulations: you have just un-personed yourself because you refused to let others help you become an actual individual in society. You're a sack of meat hooked up to a machine. ... It's probably obvious I feel strongly about this in particular.
At the end of the day, we can at least agree that people should learn to read and write? For now?
> Students taught to solve easy problems by themselves will be at a big disadvantage in the workforce compared to students taught to solve hard problems using AI.
What hard problems could students solve with AI that requires the students to be especially trained? It seems you are thinking of GPT-3 style "prompt engineering". That's a thing of the past. Students can just copy the assignment into the LLM. They don't need to be taught to do that.
There are a number of assumptions in what you say that don't necessarily hold.
1) That school is simply about landing a job.
2) That there is a value in students knowing how to have the AI do problems for them.
3) That follow-on effects of manually solving difficult problems is discountable compared to the direct output of the work.
I would say you're absolutely correct in that people pay for the result and they don't really care how you got there. But that's a pretty shallow rationale which overvalues the ability to be the conduit from the source of requirements to the final output and undervalues the individual ability to think for one's self when faced with the challenges of technological, geopolitical, or simply uncontrolled personal circumstances.
"The conduit", who you seem to be believe is the one with marketplace advantage, is exactly the person I would say is the most vulnerable. Not because getting the AI to produce demands is without value, but that its quickly becoming a task that doesn't need the intermediary at all. Those magicians that can prompt/agent/mcp/etc their way through to positive successes are actively being challenged by the very AI producers which our conduits people now depend on. Removing the need for intermediaries would be a great competitive advantage for any AI vendor able to achieve it. But insofar as intermediaries create output from LLMs, they'll not be very well differentiated: the common wisdom tends to be the output, lest the AI be accused of hallucination or being overly supportive. But when everyone is using AI for everything the opportunities will be in arbitraging that which is missed by common wisdom... filling in the cracks that any responsible AI would simply never venture to consider. Our conduit-person will be at a decided disadvantage because it takes real thought to know when it's best to color within the lines, and when it's best to not do so.
And that's really it. A good education is teaching you about the process of thought and becoming practiced at thinking. I would expect a better educated, thinking person to more easily adapt and make use of technology such as generative AI to solve problems more so than a person that just knows how to deal with today's prompting needs. The thinking person will be able to understand the bigger picture to better get a consistent and high quality series of results than the person just getting results as needed.
And that's really it. The output of a good education is you as a thoughtful & knowledgeable person: the output on the page is merely a means to that end. But if you focus solely on the answer on the page and the only important thing... you're really evaluating the AI, not the person that acted as intermediary.
In otherwords, if the person following your advice comes for a job, simply ask them which AIs they used in the interview and then just sign contracts with those vendors instead... you'll get better bang for your buck cutting out the middleman.
> Rather than banning the use of AI and trying to catch students who are cheating, why aren't they creating schoolwork that requires AI?
That is way to high recurring cost that many won't be able to afford. One could get a second hand calculator or even computer, and then additional resources needed was one's willingness.
With mandating AI usage, we'd only increase the gap of haves and have-nots. I personally do not like the idea.
I hope so too, however cheap is relative. One's ordinary morning coffee is a full day wage for someone else. If we could have decent models fitting laptops of most students, that would be point where we could possibly treat AI as we treat calculator or computer today.
Just to put things in context, https://www.bbc.com/news/articles/ce8444gex65o shares income for a good number of people now a days. (note that many of those workers are taking care of a family of 2+ members, most of the time)
I remember a TI-89 being mandatory for my AP math classes (calculus and statistics). It was utterly essential for solving problems in a reasonable amount of time. There were programs available to assist families who couldn't afford one so their children wouldn't be left behind.
AI in it‘s current phase, definitely. However, we‘ve been seeing the transformer architecture plateauing in the last couple of years. There are still improvements, but open source models are catching up.
I feel like at this point it’s an inevitability that given enough time, capable models will be cheap enough for everyone.
If poor students have capable models but rich students have much better models that go the extra mile for a great mark and do everything in a single prompt, it would still be unfair.
For it to be fair, you would not only need good free models, but actual parity between free models and the highest subscription tier the big AI companies can offer. And I don't think that will happen in the short or mid term future.
When I was in AP classes in high school, you were required to have a TI-89 calculator. If you couldn't afford one, there were assistance programs.
You were not allowed to use a TI-92, which was the next step up. It had built-in solvers for many kinds of problems.
I'm not saying this isn't a concern, but addressing financially-based inequities in learning is not a new problem within certain bounds. There's established ways to deal with it. If we can get AI cheap enough that you can cover a year of education with $100 then we're in a good range.
A critical difference between a calculator and an LLM is that a calculator doesn't make decisions. A calculator performs the operations you type in, nothing more. An LLM does make decisions. The human operator of the LLM needs to be able to evaluate the decisions made by the LLM. That requires education and experience beforehand.
An LLM is a force multiplier only, not a replacement. It's a personal assistant to an expert. To use an LLM in a acceptable way, you still first have to learn how to do what it does yourself. I think your suggestion for people to be taught how to use LLMs is justified, but they should do so only after first being taught a no-LLM curriculum. I think this should be entirely after what the notion of an education was in pre-LLM times. Don't incorporate LLMs into our current education, instead teach use of LLMs after our current education.
continuing on personal asistant analogy, i bet even now we have ultra rich people who are not smart enough to do things themselves but are smart enough to buy (hire) smart people to work for them. And even allow them to make decisions without understanding them. But with only a guard rails : does this produce wealth for me. If yes, do what you need, i don't care :)
So this i think is applicable to AI also, pay for smarter than you AI's pit them against each other, let them supervise each other and measure the outcomes you need. Who cares how they achieve that (sound clinical and scary)
It has gone well enough for these rich people because their hired assistants were human. With humans there is some degree of responsibility, liability and a slow tempo (= time to correct) involved. LLMs are a significantly different beast.
Rather than banning the use of AI and trying to catch students who are cheating, why aren't they creating schoolwork that requires AI?
It's an interesting debate, but I see several reasons not to.
1. As a human you need to learn gradually, e.g. in CS you need to learn the basics of programming before going into more complex stuff. If you embrace AI from the beginning, it can let you skip the basics (why would you code a simple 200-line program if the LLM can do it?) and then you don't have the fundamentals when you reach the more complex level where human thought is needed. It's a similar problem as firing the juniors (because AI can do their work) but then who will become senior if you have no juniors.
2. If you evaluate coursework with the expectation of students using AI, those who pay the $200 subscription will have an advantage over those who pay $20, and in turn over those who use free LLMs. The only way to make it fair would be to provide all students with the best available LLM.
3. While I have heard the analogy with calculators many times, I feel that LLMs are at a different qualitative level. The calculator doesn't really replace human thought, or if it does, it's only some very specific form of it. LLMs replace human thought in a very broad way, so I think they are much more dangerous for learning.
I agree with all these points. The algorithm to multiplying being done by the calculator, is not the same because you still have to learn why you are multiplying. If you rely on ai you may not know why the ai is giving the answer. All you’ve learned is copy paste. Figuring out how someone can use ai and give the wrong answer will solve the a huge problem. But it feels quite difficult at the time being. Ai is so general it’s hard to think of how to pose the question in a way that ai can’t answer. Maybe submitting prompts and judging prompts that bypass understanding. What if the test was more how to teach the ai? Meh ai in education is filled with gotchas.
The issue I have is that people are still trying to shove AI into a pre-AI educational paradigm. And yes, if you introduce AI into a world where people are still trying to teach kids the way that they were taught in the 20th century, AI looks like a threat because it allows kids to cheat, both testing and themselves. But we have the option to stop trying to teach kids the way that they were taught in the 20th century, which I think people here are fundamentally not understanding.
It's an old book now, but Neal Stephenson's Diamond Age includes the vision that we should have for education. We literally have the tools today to build his fictional "Young Lady's Illustrated Primer." What he envisioned was not that far off from an iPad with a Claude subscription where Claude has specific goals for the conversation. It's not teachers lecturing a class, it's individualized education where an AI teaches students at their own pace using their own interests. And built into AI is the ability for precocious kids to go beyond the curriculum, either on tangents or to more advanced subjects. This is impossible in a world where a teacher is trying to shepherd dozens of students through a curriculum as a group.
In the 2010s, we got some of the way there with Khan Academy. It was genuinely new that a student could rewatch something until it clicked rather than having to digest a lecture and have any question that didn't immediately spring to mind go unanswered. AI offers the possibility to go a step beyond this. Instead of rewatching the exact same content, AI can present it to the student in multiple ways based on a student's confusion and keep explaining it until a topic clicks. It can find examples of things that a student finds interesting to show how what they're learning isn't just theoretical. If a student likes space, the AI can discuss how the trig concepts they're learning apply to the Artemis II mission. If they like sports, it could apply the same concepts to tennis. Students in literary classes could read different books according to their interests while AI ensures that they understand the same sorts of concepts while discussing them. By customizing based on the specific curiosity of the student, it can make learning far more engaging and actually fun.
To address your #2, schools should be working with Anthropic, OpenAI and Google to shape a new personalized paradigm of educating students. They should be working out deals that give access to AI to their entire district. If I were heading the Department of Education, I would go a step further and get companies to bid on a contract to put their AI in the hands of every public school child in America. A version of the AI where teachers input their curriculum into the AI and students work through it with the AI, either alone or in small groups and the AI reports back to teachers so they can intervene where they are most needed would allow school districts with staffing shortages to serve more students more efficiently and with better results.
Sometimes it feels like our current system of education is only secondarily concerned with students actually learning and the primary concern is testing students to sort them into different tiers to be absorbed into different strata of our workforce. AI does compromise this sorting process to some extent. But if we can get back to the true mission of education and think creatively to deploy AI to best educate students, we have the potential to transform education like never before. What if we don't need to test students? An AI can give an individualized assessment of how well a student has grasped what they're supposed to be learning based on weeks of individualized work. It's as if we can give every student their own private tutor who will report back to the teacher on the student's actual progress. When you have that, stress-inducing exams are a ridiculous substitute.
I've been pretty shocked at how closed-minded the responses to my comment have been. We're supposed to be a community that envisions radically better futures that can be built with technology. And here we have a revolutionary new technology that upends a staid and increasingly problematic part of our society and the majority of the responses are geared towards explaining why that staid and problematic institution should be maintained unchanged. AI is fundamentally a danger to our current education model, but that model can change radically for the better. And I would've hoped that more people here would have recognized that.
Have you tried to learn a new topic with ai? Not like, learn the answer to a fleeting question. An actual large topic one might take even a short course on.
I say because most of the experiments I have seen in this space have failed. Chatgpt education was quietly removed not that long ago. Khan academy recently said that their Khanmigo AI tutor was facing challenges because students don't want to use it. Its a long-standing observation in the field of education that the miracle of the internet, computers being ubiqutous, phones etc hasn't clearly resulted in improvements to education (theres some minor evidence for or against it, but no blindingly obvious effect)
I worked at one point in ed-tech and the longer I was there the more I realised that nobody wanted this. Students only used it if they were made to, teachers only did it if admins wanted, admins only did it if they were sold on it, and the sales people seemed to be the only people who actually thought it was helping anyone
People often seem to think that teaching is the process of neutrally presenting facts to a learner. That the better and more clear the facts are presented, the better. But any book can do that. The entire business of education, as it were, is the management of motivation in students. Exams function just as much as a tool in this as the institution of coming to a classroom at all, rather than sitting at home remote learning. You need a clear mind, distraction free - just bored enough to find the content acceptably appealing (content cannot be made more interesting).
The best things for education seem to be:
- getting enough sleep
- getting a good diet
- good exercise
- being around other people you respect who are also learning the content
- being around other people you respect who have already learned the content, and who you want to emulate
- having someone who visibly cares if you learn the content or not, who expresses and reinforces that they expect you to learn it. (both "you need to focus, stop messing around" but also "you are capable of this, its hard but I know you personally can do it")
- having a good reason to want to learn the content
- having time pressure to learn the content
note that literally none of this regards the actual presentation of the content. Books have existed for centuries. A well motivated learner with all of the above will find the content. That is in no way the problem
An argument might be "but why, given all of this, are teachers actually teaching, why arent they standing up and lecturing soley about the importance of learning the content" and to some extent thats fair (although lecturing does satisfy some of the above if you think about it), but they ALSO tend to assign reading, link to further resources etc. Probably half or more of teachers speaking time is given over to procedural stuff working towards the above goals. Explaining the process of upcoming exams, worksheets, homeworks that need to be done (time pressure), demonstrating their love and knowledge of the subject (people you respect who know the content), building rapport with students (people you respect), holding people to account/motivating, explaining why the content is important, and trying to build good habits in their students (organising study groups, project work, dealing with problems and creating distraction free conditions)
AI are a poor facsimile of the above. Though they may try to replicate it, their inherent lack of physicality and humanity makes much of the power lose its effect. I dont care if a chatbot is disappointed in me. I'm not inspired by a chatbot that claims enthusiasm about a subject. I don't care if a chatbot tells me I absolutely must do this by next week else ill be left behind (behind who?).
I think the controversial part about AI is far above the "it is a new technology so use it".
When the scientific calculator was invented, people could easily know what went into its production. As in what circuitry appears in them. You knew that if you bought it, it is yours. Want to program it? Grab a book and do this. The whole package would be a fixed price. You are in control.
With AI? You are not at all in control. You rely on a big tech giant (or just like 4 useful ones) who is riding what people controversially still call an economic disaster. You are relying on a technology that is designed to very likely bait-and-switch you. As soon as you get too comfortable with AI, the big tech companies can just bump the prices up and you will not be able to say no.
You rely on a technology that you do not control.
The comparison of AI to a calculator or any other technological advancement for students is apples and oranges for that reason.
Imagine giving a student a personal AI datacenter to carry with them. This may be more of a fair comparison.
PS Training students on using AI, especially for free, is setting them up for reliance on the big tech companies and the subscription model.
Your comment is framed like "giving a student a personal AI datacenter to carry with them" is unrealistic, but in fact it is easy for anyone with access to $1000-$2000 worth of compute to download and operate exactly that for free, with performance perhaps a year behind the state of the art.
There are simpler LLMs that run on much cheaper devices and are still helpful for baseline tasks. Of course they are prone to hallucinating once they reach the limits of their world knowledge, but this also changes their effectiveness in an educational context: they can help you polish a paper (much of their reliable knowledge is about language, syntax and style/pragmatics of the input texts), but you still have to plan the writing on your own.
Maybe teaching students to take their whatever devices to run AI is the way, sure. All I tried to say is if we're teaching students to think independently, we should teach them independent tools.
That however requires significant investments - either each computer gets a powerful GPU for local inference (which cost a fortune) or the school gets a rack worth of compute. Most schools however even struggle to get their children fed.
Another issue is that it forces kids to stay in school for longer to do their homework, which can be a serious problem in rural areas where public transport is limited, so parents are forced to fetch their kids from school which may not be compatible with working hours.
Replace "using AI" with "asking your parents". From a student's perspective, their parents are probably an expert in anything, but sometimes might make things up and they won't be any wiser to notice, because they don't yet have the basic knowledge to know what to double check for. Just like LLMs.
Why doesn't the essay class allow us asking your parents to write it for them? The art class, why not ask your parents to paint something for you? Geography, why not let ask your parents during a test?
The difference is you learned useful maths teaching your calculator. At the moment a teacher can't tell if you even read the LLM output. Even if all future literature is writen with an LLM it is highly likely the are skills you need to learn to become a best selling author.
- made tasks easy that were a necessary prerequisite for advanced math (basic arithmetic), but not what the lesson was supposed to be about
- could in theory also let students skip over what they were supposed to be learning (applying the correct operations in the correct order to solve a problem) but doing so would require programming or getting a program from someone else, which the teachers probably figured was a high-enough hurdle to accept the risk
Hence, scientific calculators helped teachers by removing unnecessary friction.
Meanwhile, current LLMs
- will happily attempt to do the student's entire homework for them
- cannot reliably be restricted in functionality to leave the part the students are supposed to do themselves to the student
Hence, LLMs undermine teachers by removing necessary effort.
Sure, in theory LLMs could enable even more focused lessons by removing even bigger unnecessary frictions (e.g. in history class, have a LLM scour a large collection of primary sources to exhaustively list passages mentioning a certain topic), but students cannot be trusted to use them this way.
Hence, teachers are trying to use all kinds of tricks to ensure that what they wanted to teach actually passed through the student's brain at some point.
> - cannot reliably be restricted in functionality to leave the part the students are supposed to do themselves to the student
Small local LLMs are essentially that. If an LLM can tell you to eat rocks as a tasty snack or use glue to make the cheese stick to your pizza, imagine what it says when you ask it to analyze/explain complex academic subjects, or solve fiddly problems. But it will still reliably help you polish your language, like a subject-specific dictionary/thesaurus.
Small LLMs are certainly restricted, but they're not reliably restricted in a way that aligns with educational goals. Even Qwen3.5-0.8B, the smallest, dumbest LLM I have lying around, can solve quadratic equations. In math class, learning about quadratic equations for the first time, such helpfulness defeats the purpose. In physics, learning about ballistic trajectories, it might be acceptable, as long as the students themselves come up with the correct equation to model the problem, and do the reasoning to identify which of the two theoretical solutions corresponds to the physical setup.
A single model that can do many things, some less reliably than others, independently of what the requirements for the lesson are, is not a good tool to give students. You would need something that can do ancillary tasks perfectly, but won't do the part student are supposed to practice doing themselves. And what exactly that is changes with every exercise.
If you want to give students access to thesaurus functionality and nothing else, you're better off with a thesaurus.
Because AI is different from calculators. People stop thinking and just excel at verbose and incomplete rationalizations of every evil, just like you do in your reply.
Yeah... We had those bulky TI Voyage 200 graphical calculators in school [1]. They could do everything the teachers could throw at us up to the point of having all but a few formulas build in.
I would say that definitely shaped me in a way where I rarely bother with the underlying details and tend to focus on how high level abstractions interact. [2]
[1] German "Mathe-LK", we could chose specializing in two things, for me it was math and computer science, the later being quite novel back in 2003.
[2] I _do_ tend to specialize in things, but e.g. for LLMs or GLMMs, while I do have the capability to understand the technical details, I just don't bother.
At this stage, I can no longer take comparing LLMs to calculators as a good faith argument. That’s a talking point, a framing device, one whose flaws have been explained ad nauseam (as exemplified by the sibling replies), and I’m left questioning either the reasoning abilities or the honesty of those still making it.
They. Are not. The same.
Have you ever known people to commit suicide, kill, or give themselves rare diseases because of their calculators? How about people dating their calculator and going batshit for a software update?
Not to mention learning to do on your own is a useful skill to teach you to think, and an essential skill to (as you suggest) verify answers. People not understanding how things work is exactly why they take bullshit output from an LLM as gospel.
I also note that such arguments tend to be profoundly selfish and self-centred. Your anecdote happened to have an outcome you enjoyed and benefitted from, but I bet that wasn’t the reality for all your colleagues. Just like you are glad for the calculators in your class, some other student may be glad for the lack of them in theirs and it may be the reason they got into their field of study.
We weren't allowed to use calculators until aged 16 and I'm glad because I learned to do mental arithmetic. Many people I know, even in STEM, aren't able to do that, or lack a feel for numbers.
It was good for you but you don’t address reality of life.
Here’s one possible scenario: After graduation, you (or somebody else) shares the program with friend, with a promise to not to share further. Soon enough, it’s on everybody’s calculator. What did real educational thing for you, is just cheat where one needs to press the right buttons and get the right answer. This completely destroys the educational purpose, but significant amount of people just don’t care and want to get a pass.
Yes, there always is a counter weapon by teachers: for example, to point to random line and ask to explain and whatever, but this is not (always) scalable.
I’ve seen this in reality in college, when there was a cs/database course final project implementation, written in Delphi (very popular at a time in xussr), that was passed from year to year, that the professors and ta were so fed up, that I got almost auto pass because I wrote mine in C++…
——
To summarize - the overinreasing amount of pure slop is seen everywhere. Regular multi-thousand line prs where author didn’t even bother to look into code, written by ai. Just prompt -> commit, push, or. Nobody wants to deal with that
Same is happening here - u it’s not to punish people who use tool in proper context, it’s to filter out people who just don’t give a fuck.
The calculator analogy comes up very often, and it’s a good one because it also illustrates where AI diverges.
The other analogy is taking a forklift to the gym. Sure you lift weights, but you don’t really do any exercise to develop your own muscles.
AI automates a significant chunk of the exercises. So you are left with people who didn’t build any mental muscles.
This would be bad enough, but it’s worse because AI severely benefits experts who have build mental reflexes/taste and can judge / verify output with minimum information.
Interesting thought! I wonder what kind of assignments, etc. would require AI that would still allow assessing how the student has absorbed the knowledge. Maybe the assignment would require sending in the prompts and the whole conversation how the problem was approached and solved?
With the calculator analogy, I think that the calculator automates executing certain algorithms (like multiplication, etc.), but using AI takes away some (most?) of the thinking.
Bad idea. This is pretty much endgame territory you are talking about.
You would give the brains of the younger generation to American tech oligarchy, a class of people openly hostile to the principles of the democratic rule of law. If you want to see the damage actors like Fox News et alii alone can do, just take a look around in the US. Now imagine them taking over the parenting and teaching role; you wouldn't need gerrymandering if you can control people's beliefs.
There’s one significant difference between: typewriters and calculators are a one-time-paymeny device. LLMs are subscription based baked by billionaire companies. That alone leaves LLMs in a bad enough place
You have a point, but it is very ironic that this sounds a lot like the old argument "you must still learn to compute by hand because you won't have a calculator with you all the time".
Just like computers in schools nowadays, I will concede that part of the education should be about learning how to use AI - but they should also learn how to do research and accomplish things without the help of AI.
>why aren't they creating schoolwork that requires AI?
Because this makes a subscription a requirement for education, and thus advances the grift that is subscriptions, rent-seeking and dependence on a service. This isn't something we should ingrain into our children from an early age.
Calculators were buy once, use forever. Subscriptions to slop generators are a long term dependency and I want my children to not be exposed to that until they can decide for themselves.
> So I can't help but wonder whether schools are going about this all wrong. Rather than banning the use of AI and trying to catch students who are cheating, why aren't they creating schoolwork that requires AI?
Back in college, my assembly class was in MIPS (incidentally taught by Professor Larus of SPIM fame). I remember slogging through writing the assembly to compute factorial and saving registers and then dealing the frame pointer and the stack pointer.
One of the other students had access to a DecStation, wrote the program in C, and ran gcc -S to get the MIPS assembly from it. However, the compiler realized the for loop (and tail call) optimizations and instead of making it a recursive function (and help us practice and understand $fp) write it with a jump instruction instead.
Aside from getting a 0 on that homework, they struggled with the next assignment that presupposed understanding of how write function calls.
---
You could argue that learning C makes needing to learn assembly irrelevant (and MIPS is even less relevant today than back in the 90s). But for learning in school, it's not about the assignment but rather the journey that one takes to get to that assignment and learn from it.
Being able to check the answers that are provided to someone requires the understanding of what goes before and beneath the answer itself.
Writing the assignment in C when you're learning the before and beneath of "this is how assembly works" means that when you later take the complier class you won't be able to debug if your code generation is incorrect.
Working with an AI as the primary tool for learning problem solving keeps the person at the higher level. There is some foundational level that a person needs to learn without relying on an AI to do it for them.
The AI and other abstractions of the underlying problem do allow us to work with more complex problems. Would you trust a bridge built by an engineer who built it with AI and didn't understand the underlying math and physics themselves?
This is especially at issue in college where some students are taking classes to get the requirement out of the way, some are taking classes for that directly - an accountant is taking the math class to do math with the numbers, while the engineer is taking the math as a prerequisite for physics which is a prerequisite for a material science class which is a perquisite for a soil mechanics class.
If you don't understand the various foundational levels without using AI then trying to identify where the AI (or any other tool) got it wrong isn't something that you're necessarily able to do.
There's a big difference between learning to program a calculator, which is deterministic, and "learning" to prompt an LLM-based AI service that is tuned deliberately to be non-deterministic and that changes every few months. I would prefer my kids learn things like thinking critically and communicating and logic, not spend their time "chatting" with an unpredictable oracle.
It's not a bad point, but the "calculator" metaphor is not the best one. First, schools don't create schoolwork. Teachers in different disciplines do. Second, calculators aren't in pervasive use across disciplines; AI products are. Third, schools provide these AI products for free - teachers don't. So a student in a psych methods class uses a free tier to create lit reviews, or whatever. The asymmetry there creates a problem for most assessments of skill, because students want expediency, and teachers assess skills and memory.
Finally, nobody seems to know what teachers are actually discussing. To your assertion:
>So I can't help but wonder whether schools are going about this all wrong. Rather than banning the use of AI and trying to catch students who are cheating, why aren't they creating schoolwork that requires AI?
Many teachers do create schoolwork that requires AI. Many teachers ban it. Everybody's trying to work out policy (to the detriment of other policy discussions, particularly the new ADA landscape). Many ED departments are captured by AI vendors - AI is a normal technological competency for ED majors at different levels. It's not that the discussion is not "are students going to work with AI?" The discussion is "how do we teach?" which is what the discussion always is.
But policy is a part of that. Admin will have guidance and policy statements, and each instructor will as well. Students, who get thrown off balance if they have two teachers with different nav bars in the CMS, want clarity of policy: can I use it? is a different question than should I use it? But "should I use it?" is the much more relevant question for instructors. The instructors passed through the 90s/00s/10s blissfully unaware of anything that was happening in these fields.
>These tools are not going to cease to exist.
Which tools, precisely? Because I'd assert your 200-dollar-or-whatever tier that runs out of tokens on Monday does not functionally exist for most students. I don't know what happens at MIT but Penn State satellite campus students aren't whipping up agentic solutions to "I have a summer online course with discussion boards." They're just plugging that shit into whatever chatbot they have on their phone. Honestly, most online courses aren't even worth that effort, but: different discussion.
The only reason there is pervasive student use is because someone made it free. The CoPilot window that comes with the basic tier of 365 and all the other in-app copilots are what my students have; the Google Docs stuff exists, and the Grammarly stuff exists, and the best way to "ban" it (which is just fine as an approach) is to make it even slightly more expensive. If someone does that, yeah, I think some of these products might cease to exist.
>I think we're doing them a disservice by trying to give them the same education that was given to those from previous generations
Students graduate into a world where all kinds of stuff is going on and all sorts of ideological forces shape what they encounter in schools. "AI" more a dominant element in that force than the calculator was because calculators didn't have hundreds of billions of dollars of investments, and I'm not sure I ever knew what the prevailing political project of Texas Instruments was. Maybe TI's CEO had a manifesto; tbf, I do remember IBM's corporate culture being strong enough to drive cultural change. But it would be naive to not recognize that these products are made by people antagonistic to the idea of education as a compulsory and public good.
> The only reason there is pervasive student use is because someone made it free. The CoPilot window that comes with the basic tier of 365 and all the other in-app copilots are what my students have; the Google Docs stuff exists, and the Grammarly stuff exists, and the best way to "ban" it (which is just fine as an approach) is to make it even slightly more expensive. If someone does that, yeah, I think some of these products might cease to exist.
So the solution is raise cost, thus making the richies have access but the poors not to?
I guess thats one way to academically bin students; just put yet another financial gatekeeping to the academic process?
Bit silly to read that as a "solution." Financial gatekeeping? Unis don't set costs for AI products.
>thus making the richies have access but the poors not to?
I don't know that you need "thus" to describe the status quo. My school has tiny little Dells that are ten years old inside our Prometheus units. You think we're gonna upgrade each unit in the school with costs as they are, or do you think MS is just going to suddenly "provide" a cloud computing solution?
Imagine military instructors would say that it's important to focus on cavalry and bayonette use just because they can't figure out how to adapt their curriculum to the new reality.
Why are people promoting the idea that exams are not written or given in person anymore? I graduated relatively recently and maybe had 1 take home exam during my entire education. Every other exam was proctored in person and written. The professor who made the take home exam also made it much more difficult than a normal exam so I would not really say it was easier than a normal in person test.
Things have changed drastically since COVID-19, at least in the US. Tons of schools and universities shifted to online systems, and never abandoned the systems they built up when it was time to go back to school.
I graduated in 2020, so I've only gotten to see the changes secondhand through friends and family who are teachers, and through my sibling who graduated a few years after me. But the difference is staggering.
I loved take home exams because they allowed me to study before hand but not have the insane pressure and condensed studying required for exams in the classroom. Even though they were normally much harder and longer, I liked them. I felt I learned much more through them because I could take the time to understand concepts I had missed without feeling the time pressure of in-person exams.
It's a shame that humans find a way to cheat ourselves out of things that benefit us by over "optimizing" the wrong things.
Exams in classroom with all the time pressure is also an important part of education. May be they should be low percentage of grade to prevent too much stress but it's am important learning experience
I disagree. Take home exams represent how work and progress occurs in the "real" world. There's nothing in the post education world that resembles in-person exams.
Maybe the medical profession is a counter example.
> There's nothing in the post education world that resembles in-person exams.
I’d argue that dealing with any high criticality operational incident is like an in person exam (maybe even the most difficult kind, the open book one) if you are the one responsible for fixing it. Everyone is looking at you, you have time pressure to solve it ASAP and you can’t afford the time to dig through all the docs on the spot. So there’s at least some similarity with some real life situations.
I can understand that argument, but I don't think in-class exams are specifically good at that. I think for those situations, it requires more specific training regarding the domain, which is why I gave the medical profession as a counterexample.
I'd like to see some data on this. My general-ed recall is minimal, and in programming before school, I certainly learned a ton more by coding than by testing. That's my perception of my time in school, as well.
Take home exams were very common when I was in school, which was before you could get answers on the internet. After internet answer and cheating sites came along, a professor would have to either not care and let cheating run rampant, or struggle to constantly make unique new kinds of take home questions somehow. AI has basically killed that option too.
A lot of people in this thread are talking about how they did in-person exams, handwritten problem sets in class, etc. This kind of thing is more challenging in the humanities, where the research paper is kind of our bread and butter. A lot of us have since turned to different kinds of assignments, but I am not ready to forego research papers in favor of blue book exams. I think there's some value in having to develop and sustain an argument in conversation with some body of literature (scholarly or otherwise), and that is not easily to replicate with in-person writing, at least at the undergrad level. (Doctoral candidates do this kind of thing all the time in qualifying exams, but that's after years of graduate school and fresh off doing nothing but reading 100+ books over the course of a few months.)
My undergrad was in humanities - we had multiple essays to submit throughout the year but they only counted for ~10% of the final grade, with the rest through an in person exam paper.
It was just expected that you had a grasp of the literature enough that you could argue off-the-cuff in the exam setting, and then you were given leeway if you didn't have exact Harvard style notations to exact date/titles of referenced material.
Student turn-in a mid-term paper. A professor chooses a certain number of points from the mid-term paper and asks for explanations of these in long-hand. Pulling questions this way doesn't seem like it would take more time than a thorough reading of the text.
Oh, but paper reading has been delegated to a drudge you wouldn't trust with pulling question, oh how inconvenient. Which is to say the problems AI introduces to education are strongly related to much of work already being made mindless before AI appeared.
Doctoral candidates do this kind of thing all the time in qualifying exams, but that's after years of graduate school and fresh off doing nothing but reading 100+ books over the course of a few months.
No, High school students can do this. Well, they get impelled to do this. They can't do this now but that's a testament to current education.
You can sometimes locally win an arms race by doing something really exotic that isn't worth the work to defeat, but this is definitely not a strategy that works if everyone adopts it.
In one of my classes the approach was the opposite, I’m expected to do Ph.D level work as an undergrad and am expected to use AI.
In a different one she just said so long as you say AI was used you’re fine to use it.
In the rest of them AI is considered cheating.
To say we have discrepancies in the rules in an understatement. No one seems to have the exact answer on how to do it. I personally feel like expecting Ph.D level work is the best method as of now, I’ve learned more by using AI to do things about my head than hard core studying for a semester.
If it’s any consolation, this problem of discrepancies in rules is very common at universities now.
I teach at two universities in Japan and occasionally give lectures on AI issues at others, and the consensus I get from the faculty and students I talk with is that there is no consensus about what to do about AI in higher education.
Education in many subjects has been based around students producing some kind of complex output: a written paper, a computer program, a business plan, a musical composition. This has been a good method because, when done well, students could learn and retain more from the process of creating such output than they would from, say, studying for and taking in-class tests. Also, the product often mirrored what the students would be doing in their future lives, so they were learning useful skills as well.
AI throws a huge spanner into that product-based pedagogy, because it allows students to short-cut the creation process and thus learn little or nothing. Also, it is no longer clear how valuable some of those product-creation skills (writing, programming, planning) will be in the years ahead.
And while the fundamental assumptions behind some widely used teaching methods are being overthrown, many educators, students, and administrators remain attached to the traditional ways. That’s not surprising, as AI is so new and advancing so rapidly that it’s very difficult to say with any confidence how education needs to change. But, in my opinion at least, it does need to change at a very fundamental level. That change won’t be easy.
It's not inherently contradictory, just like using a calculator could be considered cheating depending on the context. If you're just learning basic arithmetic, a calculator is cheating since it shortcuts the path to learning. OTOH in calculus, a calculator is necessary. You still have to have a deep understanding of the concepts and functions to succeed.
It's still a new tech so I'm not surprised a lot of teachers have different takes on it. But when it comes to education, I feel like different policies are reasonable. In some cases it's more likely to shortcut learning, and in other cases it's more likely to encourage learning. It's not entirely one or the other.
A better example might be physics and math classes. I was learned derivatives and integrals at the same time in those two classes, but the math one required we learn how it all works (using limits to understand why the derivative rules work, without using calculators, for example), while in physics we just memorized the rules and were expected to use the calculator.
Exactly, AI is the next calculator. Right now the consensus is that it just does the work for you, in my opinion that says more about us not having the right questions than actual laziness. In a world where the only questions are basic arithmetic, calculators do all the work for you. My opinion is that the future what used to be done by academics will be done by high schoolers and new academics will be producing work at a rate no one could’ve ever predicted.
For example, the professor who’s leading me in this project had a fellowship at a certain university in England and said he exclusively coded using claude code for a month straight, their purpose was to solve a vaccine for a specific disease and by using AI tools such as claude code they’re several months ahead of schedule.
I'm really not seeing how you can do PhD level work as an undergrad. You wouldn't have the foundational knowledge necessary to do PhD level work, and you have no idea how much of what you're learning is accurate.
Without going into too much detail, when I said “Ph.D level” I’m meaning active research that adds a meaningful contribution to a field. I’ll probably be posting on here in a couple months about it but I’ve been doing thousands of tests with beefy GPUs on a certain theory we have about small 9b LLMs under certain external constraints.
Am I saying I’m as knowledgeable or capable as a Ph.D right now? Absolutely not. There’s just not really a terminology that correctly describes accelerated learning and iteration by use of AI since the technology is so new. I can’t speak for others but as someone who’s a senior in my physics degree, I’ve been actually learning faster by using AI. It’s either a mental crutch or mental accelerator. The difference is in if you want it to completely do work for you or if you try to learn and follow along.
It’s a very under explored and new area right now, how higher learning is effected by using AI as a tool instead of as a cheating device, but historically, new tools like the calculator or computer have done a lot to accelerate learning once new rules are in place.
For what it is worth, no graduate student would say they are doing 'Ph.D level' research. It is called 'graduate level research' or just, you know, 'research'
Sounds like a fun project, I wish you the best. I ran a similar program (independent study that encouraged freshman/sophomore undergraduates to explore using microprocessors, at the time the EE curriculum was completely focused on analog circuit theory and ended at boolean logic) and it went well enough that it eventually became part of the official undergraduate curriculum.
It's not terribly uncommon for an undergrad to claim they're doing "PhD level work".
Undergrad research is pretty common and it's not all that hard to get your name on a paper as an undergrad. A lot of undergrads think that doing work that gets your name on a paper, equates to PhD level work.
For that specific one, it’s more of an independent project analyzing complex systems for 6 credits, I’m gonna be expected to submit a paper to arXiv on the subject with the professor as a co-author (fingers crossed). He said I can use claude code or any AI. I’m required to do X amount of hours per week and then submit a thorough report after about 2 months.
>I’ve learned more by using AI to do things about my head than hard core studying for a semester.
How do you know you actually learned, instead of being fed slop by the AI that isn't true at all? If you didn't study, then I doubt you'll really know if the AI is lying to you or not. I have to wonder if your teacher will too, sounds like they have kind of checked-out from actually teaching.
Reading all these comments, I feel like US universities are a joke.
I had to do all the exams in person. 100% of the grade was decided at the exam. Millions of people graduated this way and they are fine. No students were harmed in the process.
I agree fully. Not sure what they are on about, with "no labs??" in the replies.
You still do all the same things, and they are graded, but this doesn't affect your final grade. Instead, you need to pass a threshold to enter the exam, which is then graded.
The US isn't so amazing at this, it simply can be done better. Recognizing where you can improve and from whom you can learn is a great first step to ACTUAL improvement.
Just because those things don't contribute to your final grade doesn't mean you don't do them.
At Oxbridge, for CS we still had lab work. We still had problem sets assigned for CS and for math which were graded. We had one large CS group project in, I want to say, our second year. Humanities students were still assigned essays. It's just that none of this stuff contributed to your final degree classification which was based entirely on your exams (although if you didn't do your CS practicals you wouldn't be allowed to pass).
Obviously Oxbridge isn't exactly representative but certainly my experience showed me that the American style is not the only way of making education work.
The “it does X for you” aspect of technology is not completely without its downsides, for various values of X.
For example, take “X” to be “walking”. Do we have the technology that allows us to pretty much never have to walk? Sure. As far as I am aware, though, we do not generally favour a lifestyle of being bound to a mobility aid by choice, and in fact we have found that not walking when able in the long run creates substantial well-being issues for a human. (Now, we have found ways to alleviate some of those issues for those who aren’t able, but clearly it is not sufficient because we still walk.)
The problem is exacerbated immensely as the value of X approaches something as fundamental to one’s humanity as “thinking”.
> "It does X for you" is the point of many technologies. You still require knowledge to work around it.
When running water replaced the need to pump water out of the ground yourself, were people urged to "learn faucets"? You kind of just need to twist a knob and water comes out, right?
Maybe there was an intermediary stage where running water was slightly more complicated and there were more steps to learn, but devoting time to learning those steps would have been a waste of time, since the end goal of the system was for it to function without much input.
Not at all. It’s a tool. It can be used well and it can be used badly, the difference often being others things thought in a CS curriculum. But a well trained engineer using the tool will be more productive than that engineer doing everything by hand, so leaving that tool out of the curriculum is doing a disservice to the students
So you didn't have to do any course work? No collaboration? No labs? I'm not aware of any University that doesn't have coursework outside of online diploma mills.
In my undergrad, coursework did not count towards the grade for the module. But you earned the right to sit for the final exam by passing the courswork.
for people with medium to high anxiety problems and/or with ADHD, exams are an absolute nightmare that is not representative at all of workplace performance or ability to do good work.
It does! That’s why you can ask to be evaluated by a commission of professors.
If you don’t pass after 3 tries, commission is mandatory.
You also have a paper trail of written exams and midterms to back you up. If you keep getting good grades and failing the oral, people will find that obviously suspicious.
Honestly the only times I had any trouble in the orals were the exams where I baaaaarely passed the written. Usually oral feels like the chill easy part compared to written because you can have a back-n-forth with the professor.
> It does! That’s why you can ask to be evaluated by a commission of professors.
Still concerning from a statistical/psych fairness aspect.
There's a famous example of the Boston Symphony trying to fairly judge unseen applicants in 1952, and their results kept getting gender-skewed until they adjusted for the fact judges were reacting to the sound of shoes (e.g. high heels) when the candidate moved around behind the divider.
> That’s why you can ask to be evaluated by a commission of professors.
Ah yes, the classic "if you think the system is abusing you, you shall out yourself to the system that's abusing you if you want any chance of recourse." Because a tribunal run by the people you're lodging a complaint against can't possibly be biased.
If you don't get one job you should have - there are others - it's unfortunate but not life altering.
If 3 years into your marine biology program a professor who always teaches a mandatory course fails you because you're a woman who wears non traditional dress - you're not graduating and now there are no jobs. (And this is an example that actually happened to someone I know - not in a western country)
I think the oral exam is probably a great way to ascertain a student's ability, but let's be real, undergraduate class sizes numbers in the hundreds, for almost every first year class. I don't think it's possible to administer that. I think I would have loved to see that in my later years at university though, we still just did things by written exam + course work.
> let's be real, undergraduate class sizes numbers in the hundreds, for almost every first year class. I don't think it's possible to administer that
Our first year class was about 250 people. It was fine.
By the 4th year, class sizes were a much more manageable 30 to 50.
You get maybe 10 to 15 minutes with the professor (usually more in later years), they ask 3 questions with some followup. That’s 1 work week for the professor. And less than half the students even make it that far for every exam season (3 per school year) so you’re looking at something like 3 days of work. It’s fine.
I had a typewriter growing up and I remember thinking it was the coolest thing. I was amazed by it and tried writing several stories. Eventually my dad bought me a crappy old computer that was only really good for writing, and that was cool too. I loved that thing. It was small too, with an integrated monitor and keyboard, so it didn't take over the whole desk where I still used pencil and paper often
Imagine being able to do some writing without notifications going off every few seconds, and where you're not always one click away from a search engine and some website scientifically designed to drag your attention down a rabbit hole and keep it there
There's an entire industry of "distraction free writing devices" based mostly on that nostalgia/yearning (not to say that it isn't effective, but the effectiveness is not actually being measured :-)
For programming, the way I would build a curriculum is to force students to actually learn how to program and code first. This is simple by requiring them to write code inside the classroom by hand for all exams.
I would make this the focus for 90% of the first 2 years of their degree.
I would then have them spend 75% of their last 2 years learning how to use and program with AI. Aside from knowing how things actually work, there's no more important skill now than mastering AI.
First part is basically my degree. You can cheat as much you want, but when you have to write algorithms by hand on paper in the exam it doesn’t matter - you will have to learn. Then depending what you want to do for masters it’s basically machine learning, which mostly mathematics.
I don’t know why you’d need to teach anyone to code with LMM’s though. If you are a CS major (or any reasonably intelligent person) and can’t figure it out on your own, well, workd needs bakers and carpenters still you know?
What I am learning is mathematics, which what all CS including AI and machine learning, really is. ”Mastering AI” for me means building your own models and AI applications and undestanding linear algebra, multivariable calculus and the propability theory behind it all.
One consequence of LLM fraud at scale making remote/online tests & document submission worthless is it might act as a giant revitalizing boost for the bricks-and-mortars school systems. Suddenly having real teachers and students in room together has value again, for credibility and authenticity alone.
LLMs are also making having a public repo code portfolio be much more worthless as a sign of legitimacy
I can't compose at a typewriter the way I do with a word processor. I would have to write it out by hand first. If the typing is just transcription, I could just as well be copying from an AI doc.
If you're doing it in class anyway, and providing typewriters, you might as well provide a locked down Chromebook. Cheaper and better for composition.
Writing an extemporaneous essay from start to finish in one draft on the spot is a skill that students were expected to develop in my educational experience in the US less than 15 years ago or so. If that's vanished, I'd be glad to see it come back.
> Writing an extemporaneous essay from start to finish in one draft
Every English class I had in high school focused so strongly on how important revision was, or at a minimum having an outline to work from. While AP tests expected us to dash off essays by hand in a single go I empathize with OP around how useful a tool revision is.
I like this. Related, this semester I've been using handwritten quizzes in class. A simple change that's been one of the best things as it changed students' expectations of class prep. Kind of do the readings and sort of prep and you can coast in class. But if you need to write out quiz answers you're forced to know the material better as well as maintain the ability to express yourself.
I also use low-point bonus questions to test general knowledge (huge variation on subjects I thought everyone knew).
I’ve been typing for years since the 80s. However, even in the 90s I found any extended period of handwriting to be painful and laborious. I don’t think I could handle an instructor that insisted on handwritten long form but I’d happily accept a compromise in the form of a typewriter.
Is there really much point though? I think AI will keep improving, and there will be more and more incentive to use an AI which costs $20/month, instead of a human writer that costs $30/hour. If someone want's an article written, and if people like the AI article as much as the human one, what stops anyone everyone using AI?
The only answer I can think of is that people must believe AI writing will stay below human level for many years, but if so why?
OP is about a German learning class, so yes, there is a point. AI improving will not learn you a language faster than immersing in it until, I suppose, neural links become mainstream.
i dont think that way, ai will became better, but human-taste writing just feels different. like hand-made furniture vs factory-made furniture. they have different class.
I think the AI writing becoming better means it will appear more human rather than like better AI writing. I think the difference in feeling is similar to early attempts to generate faces with AI, which also seemed wierdly wrong in ways which were hard to describe, but now it's very hard to tell them apart.
I like open note exams (and perhaps open book exams, as you need to know the book well to know which page to look at) - it forces you to condense the material to the salient points and operationalise it to solve what would be more challenging problems than a simple recall exam.
When I see 'cheat sheets' - designed to be hidden on the back of calculators or whatever - then I see true application of human ingenuity and intellect.
My school couldn't afford typewriters in the 1980's and early 1990's.
We wrote assignments by hand using a pencil or pen.
Is that really complicated?
When I got to college and everything had to be typed I still wrote everything by hand on paper and edited with an eraser and a red pen to reorganize some sentences or paragraphs. Then I would go to the computer lab and type it in and print it out.
I sometimes wonder if the next step for expensive, elite universities will be to make exam-by-interview a thing, even for undergraduate work. Get in a room with no electronics allowed, just you and the instructor/TA/whoever, and they interview you. Pretty quickly they are going to know if you learned the material or not.
This will only work until somebody figures out how to connect an AI to the typewriter which will have some sort of MIC, and the person will start dictating into it with AI-assisted revisions. Once the dictation is over, the AI-enabled typewriter will be instructed to type the work out.
Testing and instruction should be modified to account for AI. If a student uses an Agentic AI for work, learning, research, then when test time comes, the student should be required to stand in the front of the class and teach the class what they have learned, i.e. "Teach Back" all they learned to the entire class student body and teacher. The entire class, instructor included, will also be required to participate in a Q&A session to make sure that student's learning is not just made up of memorization, e.g. restate the information learned but using different words, different scenarios, etc.
FWIW my Dad taught me how to type at 4yo on a huge Imperial typewriter. My spelling took an enormous leap in capability in a few weeks. Primary school teachers were amazed at the words I could spell correctly. (Didn't help my handwriting though which was still like intoxicated chicken scratch on a good day).
Aside from AI-proofing, IMO there is value in slow typing or writing. We have to think just a little bit more before putting ink to paper. There is also a higher cost to making mistakes.
As a kid, before my family could afford a home computer, I was determined to do something that resembled programming. I borrowed "BASIC Computer Games" (1978) by David Ahl[1] from the library and typed in several programs on a manual Olympia typewriter. More than just reading code and maybe even more than being able to easily execute it, I'm convinced this typewriter exercise forced me to really study the flow and the how of the code.
In college I had to handwrite all my exams and the pure math courses didn't allow calculators. For the latter you could realize you were doing it wrong if the answer was too complicated to write down. As others said the final was something like 30% or more of your grade.
If you need a typewriter, there's a company in New Jersey that makes them for the prison trade in a lucite housing to prevent prisoners from hiding contraband inside.
Not sure if China manufactures new machines. India supposedly still manufactures them.
College education is all kinds of backwards if the goal was teaching students. If you aren't speaking day one of German find another teacher. In this case you are cheating yourself out of an education if you use a typewriter.
We have amazing tools through screens, but insist on using them as if they were the old tools and act surprised when they lead to worse outcomes. Each student could easily have a friend in Germany they speak to in real time if people actually cared about breaking down cultural barriers. Schools have never cared about education, their main purpose is control. Control of the narrative and control over your life.
When I was a kid we were told no calculators because when you grow up and are in the real world it's not like you will have a calculator with you at all times. Fast forward to today and we all have a calculator on our phones.
I think AI should be treated the same. Who cares if it assists in a lot of the work that is a good thing. BUT as we all know AI has been incorrect on many things so I think what would be a much better learning practice would be to forget if AI wrote the paper and focus heavily on students backing up their claims with sources. So if your paper says ABC is true and AI writes it up in a perfect paragraph you would still need to confirm the facts as true and find a reputable source that shows it to be true.
If I was a professor, I would make a very clear policy they AI is not to be used on assignments, and would repeat it throughout the semester, but make no effort to actually enforce this and even make it easy to abuse.
Then, for the final exam, drop the bomb: in person, handwritten, no outside references, mostly the same assignments we’ve done before. If you fail, it’s over for you. If you stayed true and studied, it should be easy for you to pass. If you used AI all semester, you did it to yourself. Those who complain will have their past assignments audited and if AI was used they are reported for plagiarism. This will be the most valuable lesson.
Surely there's a middle ground? Get some old wordstar capable 86 class clones and leave the GUI off. It's typing on a keyboard without the confusion of the internet or clicking on glowing icons.
You could just as easily have a bunch of old desktop computers with no network interface uplink going out of the room active whatsoever, running some basic GUI and libreoffice, connected to nothing but a dumb copper ethernet switch in the same room and an old HP laserjet with a 10/100 ethernet interface in it. No need to force people to deal with typewriters.
A typewriter is extreme. In school I used an AlphaSmart (https://en.wikipedia.org/wiki/AlphaSmart, because my handwriting sucked). A laptop without internet would also work.
After 30 seconds the site shows a fullscreen popup:
> The Sentinel not only cares deeply about bringing our readers accurate and critical news, we insist all of the crucial stories we provide are available for everyone — for free.
Thank you very much for interrupting and ruining my reading experience of your article.
I don't think that complaining about things necessitates believing that you're entitled to them. I agree that complaining about things you received for free is in rather poor taste, but I don't think that it's morally wrong in the way that you seem to think it is. If an article you read for free had a pop-up ad on it, you have not been wronged in any way and do not have grounds to sue them, but you should be permitted to voice your complaint, so long as it's of the form "I don't like this" and not "look what they subjected me to, those monsters".
Just because you expect someone to owe something to you, doesn't mean they think they owe you that thing.
If someone gives away something free, they can and sometimes do wash their hands of it. That doesn't prevent you from expressing your opinion on what you think they should change about the work, but they're not under any obligation to do anything about it.
Someone made a thing available. You can take it as it is, you can make noise about what you don't like, you can make it better, or you can ignore it and move on.
If someone is providing a mix of useful and garbage information, well, take your pick from the above.
Remembering my college typewriter-use-by-quarters (coins) on a timer like being at the laundromat, I kind of love this.
At UT Arlington in the Stone Age we had a typewriter lab so folks without home computers with printers could still produce their papers typed, which was required. I had to get a roll of quarters ($10) to do a single paper. And the erase tape was always so used up it was useless.
It was one of the most sadistic things I remember about my college experience, trying to type on those crappy typewriter on a timer. With no errors. And I literally wrote it by hand before trying to transcribe it.
To me this nostalgia is pointless. AI is here and it's good enough only going to get better. The classroom should be about using AI better not ignoring it.
But that would require the teacher to be good at AI too. I think that's the problem here.
> The classroom should be about using AI better not ignoring it.
No, it shouldn’t. I’m not bearish on AI but it shouldn’t replace any part of a classroom where the objective is to learn and communicate in a new language (German). The typewriter argument is memorable and interesting - the article points out the lack of editing forces kids to slow down and think about their writing, as well as iterate through multiple drafts. It’s not a nostalgia thing, they’re not old enough to have ever used one before.
I could see an argument for adding on a new class for GenAI, agents, context engineering or what have you, but considering how behind current US curriculums already are and how quickly the AI field moves, I can only see this ending in wasted time and money: even an up to date class will be stale by the time it’s over. Kids will end up learning this anyway outside of the classroom, no use lecturing them on something they’ll already know.
You can do all the same learning with AI tools, with the added benefit of developing extra skills on top.
This notion that AI automatically reduces learning seems more born out of fear than reality. There are also different kinds of learning.
People who really want to can still go back to typewriters. Forcing it into a mandatory curriculum though is a step backwards. Just like it would have been pre AI.
Just because an AI can craft a decent Japanese text doesn’t mean I can. Just because AI can write x86 assembly also doesn’t mean I can.
You don’t give first graders a calculator because they will always have one in their pocket- they end up just inputting numbers in a magic box and not learning how to do this manually which will destroy their future mathematical education. It’s about the same with AI.
If students cheat they hurt only themselves. Make sure they understand the consequences for cheating (missing out on learning) and that's about all you can do.
Depends on your measuring stick. Cheating themselves out of an education? Yep. Cheating themselves into a credential -> job - the status / remuneration of which is almost entirely divorced from the quality of the education, being aligned rather with the name of the organization on the diploma.
Former (second-generation) college professor, here. I find it almost impossible to be cynical enough about the US education industry.
When i attended university (almost a decade ago i guess, time flies) we didn't have a single exam on the computer. All exams were on paper or oral, most were without notes too. Computer science does not require computers.
The thing is, when colleges don't test students' ability properly before issuing a credential, employers start testing job applicants' ability after they've received it.
And they'll do it with all the 'unnecessarily high stakes' and 'risk of unconscious bias' and 'not truly representative' problems that written exams have; and a bunch of extra problems too.
This is usually true, but it is also true that some classes are graded "on a curve" and so grade inflation could hurt people who are honestly doing work. Also, cheaters tend to suck all the air out of a room. For example, my I.T. instructor designed a really nice oral quiz slide-show for the entire classroom. I found it a few hours before the class, I watched it in its entirety, and then when he tried to run it live, I spoilered all the answers before any other student could answer. I wasn't strictly cheating, but I wasn't being fair to my classmates' learning process, either.
This is untrue. Students who graduate without actually absorbing knowledge as laid out in the curriculum devalue the degree when they show up in the workforce lacking that knowledge. This is part of why new grads are undesirable job candidates, there’s a chance you are paying a higher wage for someone who may not have learned anything.
Things like this are well-intentioned but idk why there aren't more teachers creating optional "side quests" like these for students that want them instead of forcing them to do things like these
optional "side quests" would allow teachers to create some standard accepted "main quest" curriculum and then just create a bunch of (even possibly "fun") "side quests" students can work on in their spare time for extra skill development
If AI can do the work, maybe the test should be more focused on what AI can’t do? This is like anyone still doing a traditional coding interview with leetcode problems just because they haven’t yet done the work to figure out what to test for in a world where Claude Code exists.
Gyms are a great example actually because tractors exist to do the economically useful work. You now optionally go to the gym to benefit from fake labor that used to be the side effect of useful work. The fake labor is now what colleges are trying to sell, and it's going to kill them.
3,000 years ago, physical labor was a component of most jobs. Today gyms are for people who can afford to attend them and don't have a day job that naturally exercises them through labor. People exercising purely for health benefits, and not because the strength benefits them in their job and in other facets of their life, is new.
Huh? The gym analogy doesn’t even make sense. People didn’t go to gyms when they were farming with oxen. Gyms are popular now precisely because tractors exist and you don’t need manual labor to farm anymore but people still need the physical exercise for their health. Society has adapted to the arrival of new life-changing technology. Our education system needs to adapt to new technology like AI too. You can probably uplevel a lot of courses and cover a lot more interesting topics than before and teach real application of things you learned aided by AI. Just like when I was doing a CS major 20 years ago, they didn’t spend too much time teaching me assembly programming beyond 1 or 2 lectures (they let me use a compiler for programming assignments!).
Maybe instead of trying to teach around the abacus, we need to teach the higher level things you can reach with MATLAB.
We're doing these students a major disservice making them live in the old world. It's our fault for being inflexible, but their world is going to be wholly different and we should just embrace that.
I’m confused about too many things being measured at once. Is Phelps banning AI to ensure her students are fit to pass terminal examination? And doing so to ensure that her class has a good pass rate, proving she is a good teacher and can keep her job? What if her cohort are particularly dumb? Is she incentivized to make it easy to pass her classes to get that A you paid so much for? Or hard or make that A worth something?
My mentor, a PhD in classics, told me it was never about outcomes and only about improvement. I suppose that answers my question. If your AI gets you an A at the start of the course and an A at the end, then, in the sense that you have not succeeded over anything, you have failed.
My impression was she just brings the typewriters into class as a one-day novelty thing per course, not that it becomes the norm for the whole semester. The goal is to give the students a taste of what the old-fashioned way is like, to get them thinking about it.
Missing the /s? Most people have a calculator in their pocket all the time now.
I remember having to split bar tabs by hand at the end of the night without a calculator. The waiter delivers a hand-written bill, and we'd calculate how much each person owed, splitting drinks individually and food equally, recalculating tax and tip to derive each person's total. We'd pool the cash in the middle of the table (with people taking their own change as they go where possible), then counting all the cash and making sure it still included sufficient tip. All after many pints. I do NOT miss those days.
That makes sense. The CX-2 calculators are a bit less like the iPad era and more like the equivalent of calc I/II classes which only let you use specific TI models versus an app on your smartphone.
It reminds me of a family friend who's a bit older and did their scuba certification using dive tables, whereas when I did my PADI, I was able to use a dive computer.
My colleagues that teach hard skills courses (like data structures and algorithms) either love AI and incorporate it into their teaching at every moment possible, or despise it in the same way graphing calculators were by high school math teachers when they were introduced nearly 30 years ago.
I teach soft skills classes to engineering students, and I'm unconcerned with students using AI. I write my problems in a way such that, if the student truly understands the assignment, prompting the AI to solve the problem and iterating on it takes a similar amount of time to doing the work themselves. AI is not very good at writing introspectively about the student. In other words, AI isn't going to be helpful when the homework question is "A fellow student comes to you asking for suggestions on how to maximize their chances at landing an internship. What advice do you give them that's immediately actionable?"
Try it, plug that into ChatGPT or your favorite LLM. It parrots the same generic tips everyone tells you, with very little on "how" do perform the action in an effective way. Read it, copy it into your advice document, get a poor grade. Try telling other students to take this advice. Note how they don't because the advice isn't actually actionable enough for them to take action.
LLMs are also not very good at the follow-up question "In a previous assignment you gave specific and actionable advice to a peer on the job search. Which of these suggestions were so good you are now doing them?" A number of students write a "Mental Gymnastics" essay, claiming they are following all their suggestions (because they think that's what the professor wants to hear) while the evidence they provide demonstrates they are not. A student asking an LLM to write the essay for them consistently produces a digital 'pat on the back'; a mental gymnastics essay that ultimately makes the student realize how unwilling they are to solve the #1 problem in their college career.
I've done away with exams wherever possible. I stick to project-heavy courses. What I've found to be far more concerning than AI use is the increasing loss of social skills and ability to cooperate within the younger generations. The number of students who would prefer to fail a class instead of talk to literally any human being is astounding.
The number of students who refuse to build soft skills, and believe that tech is truly a meritocracy where the only thing that matters is 'lines of code', there's no politics, and they won't work call or crunch or give code reviews, is also astounding.
Might be an unpopular opinion in this thread, but college was made worthless for most degrees as soon as the internet got popular and silly performative shit like this is the death knell. College is about learning how to work in an industry. I'd predict an uptick in trade schools and other hands-on work like medicine, and a continuing downturn in so-called formal education for anything white-collar, programming included. Students are customers. Businesses are going to use AI going forward. No reason to waste time on this.
Education is a nice side effect sometimes but yeah, I don't know how you could reach any other conclusion. If you're motivated to learn for learning's sake, college is an annoying slog that you know you don't need post-millenium. I literally left college early and started making money instead of spending it, because I got tired of demonstrating to my professors that I already knew everything they were teaching and that it'd be a waste of time for me to come to class.
Or maybe you chose to waste your time because you treated college as a way to get a piece paper instead of as the only time in your life when you are surrounded by experts who will spend an hour a week answering any questions you can think of.
No time wasted at all, that option is also trivially available outside of college, it's called "email". There's a whole industry in tricking new adults into believing that college is not about getting a piece of paper, it's gross, and it's avoidable. I paid off a year of unnecessary college debt in 1/4 of a year of doing real work I learned how to do in my free time. It's a trap and articles like this where colleges are working as hard they can to make education less useful prove it.
It would have been more wasted time had I continued after a single year. I went to my first year of college on the advice of my well-meaning parents who are old and like most old people thought it was still important, and yet they agreed with my decision to leave after the first year on an offer for a real six-figure job because there was nothing to learn that I hadn't or couldn't have learned on my own. At least one of my own professors also openly wondered why I was there at all.
To your second question - less than a hundred, but tens. Most people who are worth listening to publish their work and their thoughts. Email is free. Experts love to answer questions about their work, professors hate doing extra work for no extra pay. The incentives here are not confusing. How much time have I taken? Confusing question. These are real people with real passion, and they answer questions with that in mind. Professors are obligated to puke up an answer. I've gotten responses in most cases, in some I haven't. When I don't get answers it's because the targets are smart and busy. If I wanted more engagement with my random questions I'd offer money, and if I had offered money every time I'd still be below par on the money I wasted on college. If I wanted to justify it - I'd say I learned enough to validate that paying real money for another 3-6 years would have been less valuable than burning it for heat.
> At least one of my own professors also openly wondered why I was there at all.
I think you completely misunderstood this interaction.
There are 2 possible explanations.
1. You are so smart/knowledgeable that the professor thinks you are beyond college.
2. You were acting like such an arrogant know-it-all that the professor was being sarcastic.
I’ve seen #1, but I’ve seen #2 many times.
You sound like you have a huge chip on your shoulder about not having a degree. I had the same issue at one point before I went back and finished (after working as a professional developer for a while), so I recognize it.
When I did go back, I asked questions in class, I went to office hours to ask questions, and I did research projects with professors. Some back of the envelope math says it would have costs me about twice what I got out owing if I’d paid for an equal amount of time with whatever experts I could find.
My strong suspicion based on the few posts I’ve read is that your attitude is the reason you had such poor interactions with instructors.
I had excellent interactions with my instructors. I interacted with them like human beings and they understood that their limited time would be better spent with students who didn't have the same energy I did. Several professors, when asked, put me through an impromptu whiteboard quiz and said yeah, do your own thing. It's great that you participated in the process in your own way. In my case I asked if I could show up for the final tests and nothing else, because the intermediate work would have been useless, received permission, and passed.
Chip on my shoulder - no, and it's a silly label to begin with. Understanding that it's for other people who value the paper more than intrinsic understanding, yeah.
EDIT: I will concede in some way that I'm proud of not having a degree, and it does influence my thoughts on this topic. I've met some real idiots that do, and I don't consider it a serious differentiator.
Also looking up the thread - at my early jobs, I was surrounded by many people who were interested in educating me on any topic I could think of, because similarly we were all being paid for our time. The difference between that and school was the assumption that we were both motivated and capable.
1. There’s already a process for testing out of general education classes. You show up, pay $50, and pass a test. You could have saved most of your first year of tuition.
2. These classes that you blew through weren’t upper level classes. They couldn’t have been because you wouldn’t have had the prereqs to take them. If you already had some knowledge of the field and didn’t need lower level classes, you could have talked to the department about testing out of some of them.
I know you didn’t walk into an upper level class on Automata theory and come up with the proofs on the spot.
No professor would in good faith tell you to go do your own thing based on what you’re describing.
If they thought you were very smart and sincere about learning, they’d encourage you to do independent study with them, do research, work with the department to move into higher level classes, or take cross listed graduate classes.
If they thought you were kinda smart, but a huge asshole, they’d tell you to go do your own thing because they didn’t want to deal with your crap.
This is all coming from experience as someone who came into school not needing the intro classes, and someone who used to be that arrogant.
It’s expensive because in a capitalist society experts want to be well paid.
As to why you’re unlikely to have another time in your life when you can do this, that’s also because of money. Life tends to get more expensive the older you get.
That doesn't work in tech based professions. In college I took music technology. It was 2 years of my own learning and explaining how everything worked to my tutor.
I'm confused at these comments because it makes me think the commenter didn't go to college. All my classes were basically open everything except for a literal laptop. Calculators were not useful.
at my university in math exams we were only allowed to use 1 specific model of calculator, and most of the exams were answered symbolically anyways, so the calculator usually was not helpful anyways.
When I did my Computer Science degree the vast majority of courses were 50% final, 30% midterm - even programming exams were hand written, proctored by TAs in class or in the gymnasium - assignments/labs/projects were a small part of your grade but if you didn’t do them the likelihood you’d pass the term exams was pretty darn low.
We already had AI proof education.
In Spain, the whole university system was like that until like 15ish years ago. Exams were king, in most courses they were worth 80%-90%, and of course always in person.
Then we did a university reform, partly with the excuse of aligning with the rest of the EU within the Bologna process (and I say "excuse" because that's what it was, because the politicians introduced some things with that pretense that weren't like that in the rest of the EU at all, and it was perfectly possible to comply with Bologna without doing them) and partly to copy the US/UK ways. And one of the pillars of that reform was continuous assessment, and evaluating coursework.
As a consequence of this, first of all working class students were royally screwed. Because suddenly it wasn't OK to just organize yourself to prepare the exam, you had to attend lots of sessions to earn points, which put students who work at a disadvantage. And second, passing by cheating became possible, even before LLMs. People tend to forget that before everyone got access to ChatGPT, some people had access to experts (family members, or even paying someone to do the work).
Now that this kind of cheating has been democratized and everyone can do it instead of just the most privileged with access to experts or money to pay them, people act all outraged. Although pretty much nothing is being done, except for using snake oil detectors, or sometimes increasing difficulty of assignments to make them LLM-proof (with which you screw the students who actually want to learn without LLMs).
They spent years indoctrinating us (professors) in training courses on how the old exam-based ways were wrong (the "Napoleonic" model, they called it... none of them seems to entertain the thought that maybe if it had been working essentially unchanged since Napoleon it wasn't that bad, and you need solid reasons to change it beyond "this is old so let's change") and the new ways were the bee's knees. Like in the Milgram experiment, it's difficult for people to back down and acknowledge that they have been wrong, even when the solution is obvious.
I think the only defense of the new model is that it forces students to learn throughout the semester, rather than just before the exam. Which is easier and more effectively engages long term memory (like doing more rounds of spaced repetition).
I definitely could tell the difference, though most of the time I just studied full 4-7 days before the exam.
When I studied CS in Germany in the 2010s (also in the Bologna Bachelor/Master system) most courses had weekly graded assignments. But the assignments didn't count towards your grade, instead you needed to reach a certain total of points to be allowed to take the exam. The actual grade was entirely composed of midterm and final exam (pen and paper exams, no computers, no multiple choice).
It was easy to cheat on the assignments. Working on them in groups was common and sometimes encouraged. The only person you could really cheat was yourself (and a TA who had to grade one more exam)
In the UK they claimed that girls did worse on one-off exams and so the one-off exams structure favoured boys. When course had more graded coursework, girls did better.
So that was the justification used to switching to a less impactful final exam.
No idea how true that is.
We were also told learning a phonetic alphabet was better for young children learning to read than using the old ABC system.
As far as I have heard, that turned out to be based on one person's fantasy and zero evidence and has actually had negative impact on children learning to read.
That sounds like a good reason to keep the exam model, since boys are already at a strong and widening educational disadvantage compared to girls.
That would make sense if the stated goals were the actual goals.
You can still (with some difficulty) take the old style exam
Getting the feeling most education research papers are written by high school teachers lol.
The UK has a real problem with pseudoscientific nonsense invading the education system.
To my knowledge they still teach about audio/visual/kinetic learners and how you should structure the way you learn around which one you are. This has been debunked for decades.
> The UK has a real problem with pseudoscientific nonsense invading the education system.
Not just the UK, pedagogy/education is a very soft science, along with any other field that revolves around human behavior (psychology, sociology, etc...).
Using AIs in experiments and studies will be an improvement even if they do not accurately reflect human behavior, just because you don't need a harm review and you can repeat your experiments multiple times under different variables.
Yes, it does have some advantages. Apart from what you mention, another one is that it's not so consequential to e.g. sleep badly the night before an important exam. It's just that I find the disadvantages to be much greater than the advantages.
If seven days of study are sufficient to pass the class, why is so little material being taught in one semester? It sounds like the exams are far too easy.
They are cramming for 5-7 days. This is not sustainable for an entire semester.
Cramming and thus not keeping almost any of it in long-term memory or operational knowledge.
Bologna process made everything worse. Pre-Bologna degrees are more valuable in public than the Bologna ones.
> and partly to copy the US/UK ways
In the UK it's common for exams to have almost all of the weight. In my Physics degree almost all courses were entirely dependent on the exams.
Including a final exam which examined the entire four-year MPhys course known as General Problems where even getting 30% was considered a good grade!
We were told otherwise, but it might have been propaganda as well. There was a lot of blatant indoctrination to push that reform.
The lack of any real innovation or major economic development in Spain would appear to be fairly strong evidence that the "Napoleonic" model wasn't actually working. Maybe this new system was even worse but I'm baffled as to why anyone would believe that changes weren't needed. Physician, heal thyself.
Spanish STEM graduates innovate just fine. They just do it abroad, where they get paid decent salaries if their work for others, or decent investment opportunities if they choose to be entrepreneurs :) Spain has lots of issues but I don't think education of the workforce has ever been one, neither before nor after the reform (it was definitely better before, but still, it's not half bad for now... Let's see what happens with the inaction with respect to LLMs).
In my public university in Spain, we always had the option to do a single final exam instead of the continuous assessment, although very few chose it. Generally the continuous assessment was less stressful, and the material stuck better, with room to digest it rather than just cramming for the exam and forgetting it right after. Generally the default expectation was that everyone was a full time student yes, but there were proper accommodations for those that weren't.
It is definitely a lot more work for the professors though, most of my family are teachers. It's a lot of assessments and it's very rare to have funding for TAs. Some think that the extra work with worthwhile for the sake of transmitting the knowledge more effectively, but not all of them do.
Frankly, you sound a bit bitter about it from the professor's perspective, and somewhat rationalizing why it is bad for the students. But students do generally appreciate it, and yes good students too, not just cheaters. I think both good and bad students end up learning more and hating the process less.
Your comments on Bologna do resonate though, it was very confusing when I continued to study in Germany and the Netherlands. The massive reforms were supposed to be for alignment with EU, but if anything it got more misaligned. They unified all 3 and 5 year degrees into 4 year degrees, but in most of EU all degrees are 3 years now, for instance.
Regarding the parent comment, indeed, my Computer Science degree was mostly hand-written exercises and exams, and it wasn't that long ago. The degree is about fundamentals, about understanding concepts and applying them, about the tools you need to learn anything in CS afterwards. You are expected to learn most of the practical skills for building software on your own, since they are ever-changing. And I have to say, that style of education has served me very well in my career.
PS: I was also surprised to learn that most of the undergrad exams in Germany, and some in NL, are oral. I can see how that might be a disadvantage to some, but writing is also a disadvantage to others. I quite liked it, less intense than a long written exam, and I think the professor can get a much clearer understanding of the student's grasp of the subject. But again, it's a ton of work for the professor, 20-30 mins per student one-on-one, giving them your full attention, adds up quickly.
Re: oral exams in Germany
Not sure when this was supposed to be the case, but for actual universities (not meant in a deragotary way, Germany has two types of higher eduction) in hard sciences, most classes are graded on a single written exam. Both in undergrad/bachelors and masters.
Unless things have drastically changed in the last five years...
This was in Freiburg University, which is among the top in Germany and top 250 globally. Computer Science bachelors and masters, around 8 years ago, most courses had a final oral exam unless many students (roughly >25) signed up to the class. Some classes like Information Retrieval or Machine Learning had >100 students so they had a written exam, but most others were smaller and oral: Data Engineering and Databases, Cryptography, Physics Simulations for Graphics, Formal Verification Methods, Bioinformatics, Planning AI, P2P Networking... I had a couple oral exams in VU Amsterdam (masters) too but fewer and not the final exam.
I know that an oral exam might seem less serious and rigorous, but I do think the professor can get a better grasp of how much the student actually understands the subject through an interactive interview.
My networking final in high school was probably my favorite test taking experience - there was a small written portion on eg subnets, but the bulk of our grade was setting up a physical network, testing it, and leaving the room. Our teacher sabotaged three parts of our network - could be hardware, router misconfiguration, etc. when we came in we had iirc 20-ish minutes to diagnose and fix it.
The best was when she barely unscrewed one of this big DIN connectors so at quick glance it looked fine, but wasn’t fully connected.
That does sound like fun, but it would probably be even more fun. to be the instructor giving this test.
> The best was when she barely unscrewed one of this big DIN connectors so at quick glance it looked fine, but wasn’t fully connected.
That's evil haha. It's the case where you unplug and plug again everything, changing seemingly nothing, but then it works
That sounds like the CCIE.
yeah was gonna say.
sounds like some of the technical exams i'ev taken, and/or one or two job interviews
Sounds more like a gameshow than education, like IT version of Hogwarts
I agree. But it hinges a lot of how it is graded.
If it is mostly a ”show your work”/”show your reasoning” kind of grading where your width and depth of attempts are more important than success then it seems OK.
It also sounds vulnerable to bias, unless the teacher modified each project in the same ways. How did they ensure they gave each student an equally difficult challenge and didn't consciously or unconsciously make it easier and harder for different students?
Presumably they do the same thing for every student. I hope they aren’t unplugging a cable for one student while taking a sledgehammer to the next student’s router.
“You have 20 minutes to get to RadioShack and back with a new router. It’s 11 minutes one way by car. Good luck.”
"For Miss Granger's ingenuity and resourcefulness, finding and patching a zero-day vulnerability in the router firmware... I award fifty points to Gryffindor."
When I did tertiary studies in programming there wasn't AI but we did our programming exams in pencil and paper. The "beneficial" prep we had and I had since high school was using punch cards. And 24h turnaround time for compiles. That really makes you think. And you learn how to desk check even thousand line programs. Intense focus, structuring for readability (to catch typos) and simplicity (catch logic errors) helped enormously. Was not unusual to change hundred lines of code and submit knowing that it wouldn't compile but will throw up the other errors I couldn't find. Our exams would give us 4-6 attempts for clean compile AND correct output. The only space where I experience same challenge now (40+ yrs later) is embedded code. Desktops and web stuff have LSPs and dynamic reloads and interpreted code (not a thing for me when learning) with instant feedback.
Lots of skills from those old days that have been lost/ignored in the pretence of productivity.
Yeah i really valued learning to code when I didn't have the internet available, if taught me patience and deep thinking, problem decomposition and organic (brain) execution
That is great for core principles. But languages and development environments have since assumed everyone has access to then internet. Meaning more "stuff" is the solution to problems (massive standard libraries or community created ones) rather that elegent language solutions.
The internet enabled all the complexity we have today. LLMs will have a similar effect, but instead of engineers actually having to understand the system (even in it's complexity) they will just be querying the oracle to build things or solve problems.
When the oracle can't help (or maybe refuses to) is when it gets interesting.
I learned HTML/CSS from a book on a computer with no internet access. Seemed reasonable at the time but in hindsight was absolutely ridiculous.
Similar to me, I learnt some html tags through a book which was sold at newsstand, once I was at my cousin's house using her computer without internet access, then I wrote a simple html page with the Win 95 cloud wallpaper as background image. My cousin was terrified how I did that!
24 hour turnaround for compiles, expected to change a hundred lines and it fail to compile, and you get 4-6 attempts, so up to 6 days to get a thousand line program working? Now you can compile and check a thousand line program in milliseconds. It seems so unarguably more productive, why do you call it a 'pretence' of being more productive?
> Now you can compile and check a thousand line program in milliseconds.
How do you define "productive?" Lines of code written per day? Bugs fixed per man hour? Fewest reported bugs per end user?
The fastest compiler in the world won't help you find all the runtime bugs that simply wouldn't have existed in the days of punch cards, when code was written with with more care and attentiveness since there wasn't a fast edit/compile/test development loop. YMMV, of course.
I personally dislike placing a heavy emphasis on exams. Assignments/projects have been consistently the most enjoyable and rewarding parts of the courses I've taken so far in university.
It's a shame that they are also way more susceptible to cheating with AI.
> It's a shame that they are also way more susceptible to cheating with AI.
They were more prone to cheating before AI, too.
Cheating has always existed at some level, but from talking to my couple of friends who teach undergrad level courses the attitudes of students toward cheating have been changing even before AI was everywhere. They would complain about cohorts coming through where cheating was obvious and rampant, combined with administrations who started going soft on cheating because they didn’t want to lose (paying) students.
AI has taken it further, with students justifying it not as cheating but as using tools at their disposal.
I was talking to my friend about this last week and he was frustrated that several of his students had submitted papers that had all the signs of ChatGPT output, so he asked them simple questions about their papers. Most of them “couldn’t remember” what they wrote about.
It’s strange to me because when I went to college getting caught cheating was a big problem that resulted in students getting put on probationary watch and being legitimately scared of the consequences. Now at many schools cheating is routine and students push the boundaries of what they can get their classes to accept because they have no fear of any punishment. YMMV depending on the institution
Those students will go on to "cheat" in job interviews. And the sad thing is that many of them will never face any consequences because the majority of jobs at large companies don't require any actual competence. The college degree allows them to check a box on job applications but the jobs don't actually depend on anything learned in college.
> because the majority of jobs at large companies don't require any actual competence.
An interesting side effect of the AI gold rush is that companies are starting to look critically at these do-nothing email jobs where someone forwards emails around and makes slides and Notion pages.
I’ve worked with many who occupied jobs that didn’t contribute much other than organizing text and sharing it around, but they got a pass because it looked helpful enough. Now it’s a lot harder to justify those positions when management realizes that having the not-really-competent person summarize communications and documents isn’t much better than having ChatGPT do it.
Almost like most degrees may not be very useful (or worth the cost) to get in the first place.
I wonder if the companies hiring students from those universities have come to understand there’s no guarantee they learned any of the material.
The dirty secret in HR is that performance in most regular jobs doesn't really depend on having learned anything in college. Companies only require a college degree to filter the candidate pool down to a manageable level. Line managers know they'll have to train entry level hires on almost everything.
IBM used to hire software developers based on aptitude test scores regardless of formal education, then put them through an extensive internal training program. It worked fine.
IBM (among other companies) used to have multi-month internal training programs based on their specific needs. And, in many cases, the expectation was that you'd be there for 25 years or whatever afterwards.
Performance does depend on whether you are capable to overcome obstacles by yourself, learn by yourself and produce by yourself.
People who got through via cheating in college tend to be low performers in work for the exact same reasons.
Most companies expect people to be productive sooner rather than later, and not every job is completely trivial.
This is what happens when you make education a business. You end up testing credit score instead of aptitude.
The problem with exams is that everyone has a bad experience with a poorly written one. Well-written exams will have questions that test students at different levels of understanding across the whole curriculum.
So a student who only understands the basics should be able to answer most of the easy questions and students who have a deeper understanding can answer the harder ones.
Well-written exams should feel pretty fair and leave students feeling like the result they got is proportional to the effort they put into studying the material (or at least how well they personally felt they understood the material).
Exams almost filtered me out of this industry before I even got started. I later went on to be a lead developer! I went to a rural high school with a poor math curriculum. I understood the concepts, but I was slow. When I got to undergrad, my first calc professor gave us a 60 question test with 50 minutes of allotted time. He told us if we couldn't do the problems fast enough we weren't cut out for the work and it would be better if we quit now. I've never felt more inadequate in my life. It's one of my only Ws.
> a 60 question test with 50 minutes of allotted time
Is this kind of test - many short questions - a standard thing for math in your country?
My university exams were pretty much all "2-question", in 90 minutes.
The first half was an essay where you have to reproduce a lesson from the curriculum, in your own words.
The second half was "the formulas" - you have to develop one or two formulas from first principles.
I once got an A- even though I got "the formulas" half very wrong. As the teacher explained later, I simply chose the coordinate system beginning at not the same place the textbook did. And this was supposed to be a bad teacher - he actually gave Ds to almost all of us (180 people). This was a makeup exam.
What does W mean? Withdraw?
Yes. You can do it a set number of times, opting to either retake the course for a better grade or simply not take it again. Either way it shows on your transcript.
Exams are kind of like interviews, at some level they are always at least a bit artificial and sometimes very artificial. I don't really think they do a good job of proving understanding but we also don't really have anything better.
Every class. Teachers think: "We want a curve, not everyone is getting an A". This pushes the teacher to ask questions not covered in the material. If they stick to what is covered it is hard to get a decent curve. So, in the end, teachers just hand out grades based on IQ, and there isn't a point of grades.
Courses aren't supposed to be pure memorization. If you can only answer questions directly covered in the material but can't apply it to new situations, you should not get an A.
Many "straight A" high school students earn their first B (or C) with this misunderstanding.
> Teachers think: "We want a curve, not everyone is getting an A". This pushes the teacher to ask questions not covered in the material. If they stick to what is covered it is hard to get a decent curve.
You've never been a teacher.
If you can't transfer the course's material onto new, unseen questions, then you might have memorised it, but you didn't understand it. Getting an A requires understanding, not rote memorisation.
You're forgetting the a large percentage of the student body simply does not care to judge the material fairly. Any difficult exam where they receive a low grade is perceived as unfair and complained about online. Professors that test fairly are reviewed harshly online. Universities that prioritize education over graduation receive poor reviews. Top universities specifically coddle and push their students making it nearly impossible to fail.
Also way more susceptible to cheating in traditional non-AI ways. And your mark ends up depending a lot on how much time you have to invest independent of how good you are at the course material.
Assignments and projects are great for learning, but suck for evaluation.
I really appreciated classes where there was rapidly demising returns to time spent :)
Another example, lit classes where the grade is based on time limited, open book exams, hand written in "blue books"
Read the book, pay attention in class, spend 90 min writing an essay, and you are done.
is evaluation that important? ultimately if you can't do the work you're only cheating yourself in the long run...
That is the traditional view, the view of those who want to improve their own knowledge and abilities, and presumably the view of those who would like to consider the degree to be a meaningful credential.
However I suspect that there are many who 1) are more concerned about the short term outcome, 2) consider the degree/diploma to be little more than a meal ticket or arbitrary gatekeeping without any connection to learning, 3) view the work as a pointless barrier to being handed said diploma, and/or 4) don't see the value of human learning in a world where jobs are done by AI and AI systems routinely outperform humans on complex tasks.
I'm curious about the objective competence of recent graduates compared to previous cohorts, not just in tech but in basic literacy and numeracy.
A lot of Gen Z are ferociously anti-AI, but for tribal and emotional reasons, not because of a nuanced understanding - which is ironic, because the nuanced reasons for being wary of AI are much stronger than the usual talking points about "stealing art".
Being tribal and emotional is going to make Gen Z easier to replace, because nuanced strategic insight is less common and more useful.
You say that but I was a Class of 2013, aka during the massive hiring boom of the teens. I tutored a friend of mine with a Ds get Degrees mentality who eventually graduated and now works an ass-in-seat job for Booz Allen or one of those types. I used to joke about it with another friend, that his diploma ought to include an asterisk and a half dozen other names for how much we ultimately did on his grades take homes. I’m pretty sure he makes about the same as me by now purely on tenure.
Personally, I dropped out despite a full ride+ becuase why would I put in work for a no name state school when I already has an FTE job as a developer out of high school anyway.
Turns out fraudulent action can still get the bag.
I agree with your premise about why accurate evaluation matters, but your post comes across as pretty bitter. Unless you’re at the job with him, you really don’t know that it’s a “I just need to show up” job he has at Booz Allen. Perhaps he has other great traits like a high social or emotional intelligence that make him good at his job beyond whatever was being evaluated on those projects you helped him with.
I think you're missing the point. The majority of jobs at companies like Booz Allen are sort of like Kabuki theater and don't require any technical competence. The main responsibilities are to show up on time and present a certain image to customers.
I think you're missing the point because PP addressed that concern.
Part of the purpose for evaluation is to provide feedback. I'm not going to claim that the form of feedback is great, but it does offer motivation to improve.
The other thing that feedback feeds into is credentials. I realize that some people are dismissive of this aspect of the degree, but it is important to pursue further studies or secure a job. While you can argue that these people are only cheating themselves, and some of them are cheating themselves, a great many will continue to cheat as they advance in academia or the workforce. In other words, they are cheating others out of opportunities.
It’s important to signaling to employers you have obtained skills useful to them.
And for most students that’s all they really care about.
If the companies stop valuing the diplomas, students will stop paying tuition to attend, and the universities eventually collapse.
This assumes employers will still exist.
You can imagine a world where the Corporate or State AI handles education, tailors it to individual student levels and talents, and assigns work based on its own direct experience of a lifetime of interaction.
You can also imagine that in that world where most humans would be redundant - unless the AI was optimising for human-to-human jobs and for evidence of unusual insight and creativity, not managing bullshit work for corporate profit.
You're telling me the path to an A is to hook up the Teacher AI to a Student AI.
The problem is that Universities are no longer institutions where people who want to can go to educate themselves but have instead been turned into the first part of your job interview. In this world, Universities are judged on how well they evaluate their students.
Yes. I care that the work I've done and what I've learned is actually good and correct. Vibes-based learning/anything is valueless.
I went to college as a MechE so unsure if compsci was different. But overall, all the “fun” projects were labs. We have three semesters of hell and all 3 semesters had 2-3 labs, and we write 20 pages or so for EACH lab a week (usually a team of 2-3).
I also think assignments and projects work best, but only for the percentage of students that are self-motivated, curious, interested, etc.
Unfortunately a lot aren't, they feel like they have to be there or these courses are the only path for them to get a good job. And unfortunately they end up in the workforce, too. You'll often see teams with one good developer and a lot of hangers-on.
Then I suppose we can go back to having computer labs that can only access white listed domains and other study materials. Students code there to ensure no cheating.
The labs I was in weren't connected to the Internet at all, only a local intranet. Though, they were all running pre-oracle solaris if memory serves, so I'm probably dating myself a bit.
Will the students have to go through security screening for personal devices?
That was common when I studied computer science in the nineties as well. Hand written exams mostly.
Writing papers is a useful skill to have. And many students aren't very good at that. I taught some classes during my Ph. D. and supervised some students with their master thesis and PhD thesis work. Many students get their degrees without that really getting addressed. At least Computer science degrees in the Netherlands just spend very little time on writing skills. You get students with high school levels of English and Dutch and that's it.
I learned to write properly only when I started my Ph. D. My supervisor made me do it right before he allowed me to submit papers for publication.
AI might actually be good for education long term. It will result in a more personalized approach, which I think is good. There are plenty of ways to test students that are more engaging and interesting for both teachers and students than some of the old ways. You can't fake knowledge when you do a verbal test. Or test people with a good old written exam.
And of course for teachers, you can automate a lot of the verification work. This can be a lot of work.
Yeah none of the problems with AI in education are new; some schools (or news articles) are just panicking because they gave their students laptops (and/or made them a mandatory part) and now the genie is out of the bottle.
But there were already heaps of problems with tech in education before AI.
My CS projects were often pretty free-form so in theory I could've just used AI - today, anyway. But a big part of the grade was a face to face interview where you actually had to talk about the code you wrote. Anyone lifting along with other people who didn't actually do any work would fall through then.
>even programming exams were hand written
Which strikes me as a terrible way to teach and test programming skills. If you're teaching to program without so much as syntax highlighting, you're not preparing your students for anything that even remotely resembles the industry they aspire to work in.
Honestly, these days universities should probably find a way to incorporate AI into their teaching, rather than fight it. Anything else is betting that AI will not stick around, which strikes me as a hopelessly naïve bet. Especially for software development.
I don't pretend to have all the answers, I don't know how to teach systems thinking in a appropriate way either. But I'm pretty sure typewriters isn't it, unless your students are hoping to get hired by Ada Lovelace, it's just not going to be relevant.
> remotely like the industry they'll work in
Education isn't any churning out a cog fitted to a bigger gear. A college education should not be preparing you for intellij or familiarizing you with va code.
It should be teaching foundational skills with the discipline. And I just don't understand how you look at the amazing things we've accomplished since ww2 and say "the education system that taught this fucking sucks"
What I don't understand is how you can look at all the progress we've made since ww2 and decide we don't have to update our education to keep it relevant, and argue in favour of teaching tech that's been obsolete for 80 years.
Students today aren't studying to repeat the progress made since the forties, that's been done by their grandparents, they're looking to drive the next 40 years of innovation, and if that involves typewriters I'll eat my hat
That way of teaching got us to the moon, created transistors, produced the internet, smartphones, quantum computers, the very AI that everyone is talking about, vaccines, sent probes into space, cured diseases, fed millions, jumbo jets, basically every single thing that society has come to depend on.
IMO, it's a bit disingenuous in 2026 to argue that technical college degrees exist for the sake of pure education and that we're best suited to continue pretending as such.
The vast majority of students and employers treat them as vocational certificates in practice, and the profession would almost certainly benefit from adapting curricula to more closely match that reality.
Foundational concepts are still necessary, but I don't buy the argument that we should continue teaching like it's 1946.
Yeah exactly, I remember having to write Java and C++ by hand in college in the early 2000s. It was also a good test how well you knew the syntax.
Syntax seems like a stupid thing to test in university level courses. That's trade school stuff. And I don't mean that as a criticism of trade schools, they just have a different focus.
Syntax is not the focus of your testing, but it’s often a pre-requisite to be clearly and accurately speaking the same language. Think not of taking off points for missing a semicolon but instead understanding the difference between the syntax for a method call and a property access. The different syntax conveys different meaning and so we should expect some basic level of accuracy to the language in question. At least that’s how I see it.
Marking at scale is hard to maintain that consistency though. It’s not whether the exam writer sees it that way, it’s whether the markers understand intent and objective over pedantic nuance
OK I'll bite. What's the difference between a method call and a property access? And why would that matter for pedagogical purposes?
Syntax is a stupid thing to test in trade school. That's compiler stuff.
Knowledge is built on foundations. Knowing syntax in one language is necessary to be able to do anything practical, which interacts with theory. You build valuable schema of the world by iterative theory and practice.
Nah. Syntax is trivial and irrelevant for teaching CS theory.
Do you think spelling is trivial and irrelevant in a foreign language class?
That's a poor comparison given they said CS theory.
I could easily imagine a CS theory course that doesn't involve any programming language at all.
Yes, there are some university CS courses on topics like discrete mathematics, theory of computation, and ethics that don't involve any programming.
Gotta know your math syntax in those classes
In 2026? Absolutely.
Would you say the same thing for teaching math or music?
Lambda calculus?!
College exam tests C++ syntax and people frame it as a good thing. Interesting. It sounds almost indefensible for me. If it were Scheme perhaps it would be okay, but C++ and Java...
We are at a point in AI where colleges teaching computer science are having students hand write code. It is what it is. This is about academic honesty^H^H^H^H^H^H^H^H integrity. In order to measure academic success we have to measure the person learning, not the machinery.
When I was in university we had the "Honor System"—no proctors, but you had to sign a statement on every exam and assignment that you did not cheat. And if you did cheat, one of your fellow students could report you to the Honor Board. Basically using the prisoners' dilemma to enforce honesty. And the Honor Board were always threatening to bring proctors back if cheating continued.
But yeah, everything was hand-written. On sheets of paper with pencil. I even had to write x86 assembly out by hand for my CPU architecture class. Of course, laptops were available back then but not cellphones and certainly not LLMs, so cheating by electronic means probably presents a stickier wicket now than it did back then.
The prisoner's dilemma requires that both people are facing the same consequences and are promised leniency if they give up the other person. I'm struggling to figure out how this conforms to a person giving up a cheater to the honor board. What am I missing?
I’m guessing we’re similar vintage. My CS classes were like this as well.
The only exception is that when I got into grad level classes we did have some big programming projects. But most of that programming happened on sparc stations, and it was actually just easier and more productive to sit at the machine in person with its nice big (at the time) display with all the other folks doing programming projects. Those machines had the standard dev toolchains provisioned that weren’t easy (at the time) to do on a dorm room Mac or windows computer.
I really think a lot of the ways we can reduce reliance on AI for thinking is to just set up systems where it’s not an inviting or rewarding option.
Same. Coding by hand was really, really hard at first, I remember. Then one day it just clicked and it was like I could think so clearly in my head, without making mistakes, and write code that would have compiled by hand during an exam. We would just practice and work for it.
I don't want to be polemic, but I really miss those days.
Yeah writing out code by hand made me slow down and think more.
One thing I recall is that the grading policy made it very clear that minor syntax issues were inconsequential in handwritten answers. And more advanced classes only wanted pseudocode. Which are exactly the right priorities.
Not only that, the tabletification of education has actually been shown (via standardized testing) to make our kids dumber. This is the first time a new generation has scored worse than their parents. Technology has its place, we need to pick and choose where.
Today just teachers walking around during an exam instead of browsing on their phone would do wonders…
Half related: reminds me of my physics teacher's test of how observant we were. The extra credit question on the test was "what is your teacher's favorite color?", which she had so far given no indication of. But while watching us she was walking all over the room in every possible direction, because the answer was on a piece of paper taped to her back.
My dad had a class in medical school where the professor dipped his finger in a beaker, talked about mechanisms for sugar in diabetic patients’ urine, and then proceeded to stick his finger in his mouth. He had all of the students do the same, who noted that the sugar was apparent.
He concluded the class by talking about the importance of observing patients, and pointed out that he had tasted a different finger than the one he had put in the beaker.
Sounds like something Restivo would do.
My AI assistant is configured to monitor the face-side camera feed and switch windows when someone standing enters the frame.
You mean an AI proof grading system. Grades have very little to do with actual learning and are there solely for signalling. It would be better if Universities shifted their focus to learning and eliminated grades/exams all together. They should seek to stop trying to be the gatekeepers for white collar work and instead focus on learning and research.
My local college used to have a test center you would physically go into. Now they no longer let teachers send their students there (idk if its just for programming or what) for whatever reason. I know they've wanted to cut down on wasted paper for years.
If my college is doing this, I cannot imagine how many others are also impeding on their entire goal: education.
Writing programs by hand is something I had to do too. Compete waste of time
Profs could now scan / ocr / grade it with AI though.
I'm sure they had some kind of submit your code as assignment and using testing as a way to grade the assignments.
Seems like anybody could just study 6 hours a day for the last month before the final, last 2 weeks before the mid term and use AI for everything else.
An old friend told me she'd hand in her assignments on paper, and would receive the computers output the following week.
Apparently you learn to double check your work!
Bonus point: even if you use AI to prepare the submission, copying it down by hand will at least force you to _read_ it.
Some people spend ages writing a cheat sheet with the intent of cheating on the test but realise that because they wrote it down and / or tried to summarise it they actually learned the materials.
Oh yeah. That was me during my philosophy class. My cheat sheet had an index before I was finished.
I remember those 6 page essays i wrote by hand at high school. Poor teachers trying to decipher my writing
[dead]
[dead]
I used to make my classes 60-80% project work, 40-80% quizzes all online.
I now do 50% project work, 50% in person quizzes, pencil on paper on page of notes.
I'm increasingly going to paper-driven workflows as well, becoming an expert with the department printer, printing computer science papers for students to read and annotate in class, etc.
Ironically, the traditional bureaucratic lag in university might actually help: we still have a lot of infrastructure for this sort of thing, and university degrees may actually signal competence-beyond-ai-prompting in the future.
We'll see.
I always preferred the "you get some grades along the way to gauge your progress but the lion's share of the weight went to the proctored exams" method unless the lion's share of the normal work was also proctored anyways (at which point it doesn't really matter how it's done).
The reason was less for myself and more because anything group related suddenly shot up in quality when the other individual work classmates were graded on couldn't be fudged.
The things I don’t like about putting too much weight in the exams are:
* It’s sort of unnecessarily high stakes for the students; a couple hours to determine your grade for many hours of studying.
* It’s pretty artificial in general; in “real life” you have the ability to go around online and look for sources. This puts a pretty low ceiling on the level of complexity you can actually throw at them.
Exams happen all the time in real life. Or rather, situations where you can't just look up fundamental knowledge. Job interviews, presentations, even mundane work tasks - all these require you to know the basics quickly "The basics" are relative, of course, but I often point out to my students: "you don't care if your doctor needs to look up the specific interactions of your various meds. You do care if you see them googling 'what is an appendix'." Proctored, in-person exams are the only reliable mechanism we have for ascertaining if a specific individual has mastered key fundamentals and can answer relevant questions about them in a relatively timely fashion. Everything else is details and thresholds - how fast do you need to be able to recall, how deep, what details are fundamental. From there, I think it's fine to hate poorly made exams, and it's a given that many folks making exams have no idea what they're doing (or don't have the resources to do it right). But the premise of an exam is not completely divorced from reality.
I think many of us would agree that job interviews (in tech at least) are horribly broken, because they don't do a good job of testing candidates' ability to do the actual work they'll be doing day-to-day. So saying exams are like job interviews is not a positive for exams. And, for most people, the ideal is to find a job and stick with it for years, so it's not like job interviews are common, everyday occurrences.
For presentations, usually you spend a lot of time preparing for them (similar to exams), building a slide deck or pages of notes that you refer to while giving the talk (not similar to exams). Sure, you do have to be able to think on your feet, but I don't think the comparison to a sit-down exam is all that apt.
For mundane work tasks, you have the internet and whatever reference materials you want (including LLMs, these days); this sort of thing is so different from a sit-down exam that it's almost comical that you'd try to equate the two.
I'm not saying I know of a better way to evaluate learning than proctored, in-person exams, but suggesting that sort of situation is particularly relevant to real life... no, no way.
Having been both a data analyst and software engineer I agree. The data analyst one? Here is 50K of Excel rows with all kinds of weirdness in it, you're data analyst right? You have 4 hours to analyze this data. Go!
The software engineer one: here is a takehome assignment. One week later: finished!
To be fair, they both represented pretty well what work I'm going to do. The data analyst didn't show that well how much I'd also be data engineering, but whatever, I was a SWE before having a DA stint. Back to SWE again though.
>you don't care if your doctor needs to look up the specific interactions of your various meds. You do care if you see them googling 'what is an appendix'.
Sounds like you're saying that it's acceptable to be a little foggy about the limits of your knowledge, as long as you remember the core foundations. For a first year medicine student, the edge of their knowledge will include things that are the core foundation of a practicing doctor. Why should such a student be tested as if he already had several years of familiarity with the subject when this is all relatively new material to him?
I don’t completely disagree with this and I do appreciate the well thought out response. I should say that most of my recent relevant experience was as a grad student teaching assistant so I’m neither just somebody grouching that I didn’t like taking exams, nor somebody with deep experience setting the things up. Some respectful quibbles, then, since I gather you have more direct experience than me.
> Proctored, in-person exams are the only reliable mechanism we have for ascertaining if a specific individual has mastered key fundamentals and can answer relevant questions about them in a relatively timely fashion. Everything else is details and thresholds - how fast do you need to be able to recall, how deep, what details are fundamental.
I don’t think this is how people actually engage with exams. I had a lot of folks in office hours who treat the exam as the ceiling of their competence, rather than the floor, and do things like cram or try to figure out exactly what topics will be on the test to study just those. If the goal is to establish a 100% solid foundation for things you have to know to be a professional (which I think is a great goal), I prefer something like Mastery Learning to the conventional exam process. (Maybe we could call Mastery Learning conventional exams a different set of thresholds, unusual thresholds if we want to look at it that way).
> From there, I think it's fine to hate poorly made exams, and it's a given that many folks making exams have no idea what they're doing (or don't have the resources to do it right). But the premise of an exam is not completely divorced from reality.
I worked with some professors who I thought gave good exams, some who gave less good ones, so I don’t think the premise is completely divorced from reality. But it seems more like something the good instructors overcame, rather than a construct that is really helpful.
I think it's all about speed. In "real life" everything can be looked up, but exam optimizes to not even having to look it up. Then any research becomes much faster.
Whether it's good or bad I don't know, I think US higher education focuses too much on ability to produce huge amounts of mediocre work, but that's the idea behind exams.
One of the reasons I've always encouraged software people to learn to touch type has nothing to do with typing speed - it's about reducing/eliminating the cognitive load of typing, you want to be thinking in expressions (sentences) not letters. (The increase in effectiveness comes from not getting distracted by the mechanics of typing...)
In real life you need to know the options and their trade-offs to solve a given problem. You don't need to know all the techniques perfectly, but you do need to be able to characterize them and compare them, from rote memory.
I agree, I think many people who rail against exams underestimate how important memory is to more complicated skills. How can you debug a complex application if you have to keep looking up every operator and keyword in the language you're using? It'd be like trying to interpret poetry in a foreign language but you have to look up every single noun. I'm not saying people can't do it, but it's tedious, slow, and you probably wouldn't think of them as a "professional worth paying for their service". Some amount of memorization is key.
It still doesn't feel to me that those things are similar. A sit-down exam is a time-limited, high-pressure situation where you're expected to demonstrate proficiency in the things you've learned over the past several months. Sure, much of that learning builds on stuff you've learned previously, but the focus is on the prior semester (or half-semeter, for mid-terms).
When I sit down to debug a complex application, I'm drawing my prior 25+ years of experience. While I certainly would rather fix the problem faster rather than slower, I don't have a time limit, and usually taking my time (or even leaving the problem alone for hours or days) can be more effective than trying to work quickly and get everything done immediate.
The last time I sat for an exam was in 2003, and I honestly have not experienced anything in life since then that feels like that. Even job interviews have not felt similar enough to me to evoke that same feeling. (Frankly, I've enjoyed most job interviews; I don't think I've ever enjoyed an exam.) That's just my experience, of course, but I don't feel like an outlier.
> It’s pretty artificial in general; in “real life” you have the ability to go around online and look for sources.
Sort of. In real life, you are expected to have immediate knowledge of your field and (in some environments) be able to perform under pressure. I'm not going to pretend the curriculum is a perfect match for what people should know, but it does provide a common baseline to be able to have a common point of reference when communicating with colleagues. I would suggest the most artificial thing about exams is the format.
> It’s sort of unnecessarily high stakes for the students; a couple hours to determine your grade for many hours of studying.
I don't like dismissing the ordeal of people who face test anxiety, but tests are not really high stakes. There is a potential that a person will have to repeat a course if it is a requirement for their degree. At least at the institutions I attended, the grade distribution across exams and assignments, combined with a late drop date, meant that failing a course was only an option if you choose it to be. A student may be forced to face some realities about their dedication/priorities, work habits, time management, interests, abilities, etc.. It may force a student to make some hard decisions about where they want their life to lead, but it does not bar them from success in life. And those are the worse case scenarios. A more typical scenario is that you end up with a lower GPA.
>There is a potential that a person will have to repeat a course if it is a requirement for their degree.
How is that not high stakes? The result of several months worth of effort hinges on what they do during a 2-3-hour window. If you had to build something and the last step involved a procedure that could potentially tear down everything you made, you would try to find a way to redesign it so that didn't happen, or to limit the scope of the damage.
If you want an easy, stress free job, sure, nothing is high stakes. You will never be invited to meetings with executives, and you can google everything all the time.
If you, on the other hand, attend meetings, you will need both deep knowledge (requires many hours of styding) and fast thinking, and the questions will make you realize you know little about many things. You need to have general knowledge of many things, not just whatever you are building at the moment. If you are successful on these meetings, your salary can and will grow.
Also, some meetings will make you realize that college and/or university are life in easy mode. Very few things you consider hard in college will be work-level hard.
This is where the alternative of a course with the other (still monitored for graded activities) option comes in. The downside of that tends to force in person synchronous rather than custom scheduling of regular tests.
The point is more about whether the graded work is actively reviewed than which individual choice is ideal or not though. Whether it's electronic or written, remote or in person, weighted towards exams vs continuous are all orthogonal debates to the problem of cheating/falsely claiming work.
I had attended a few courses over a decade ago and just completed a degree recently. The methods of cheating have changed, but not because of pencils vs keyboards.
High stakes artificial exams can help prepare you for artificial stakes at job interviews where you need to crank out a working solution in 30 mins with jet lag and someone looking over your shoulder
That's true. They do better-prepare an applicant for a job that filters on a person's ability to accomplish arbitrary things in a vacuum that is completely disconnected from the real world.
That's probably a good thing to filter on for, say, the navigation role on all kinds of crafts (from land to sea to space). There are naval roles where navigating with a sextant and memory is an important skill to have, and to test for.
But that operating-in-a-vacuum skill doesn't relate well to roles that don't need to exist in a vacuum. In most of the jobs in the real world, we get to use tools -- and when the tools go out to lunch, we don't revert to the Old Ways.
When an accountant's computer dies, they don't transition back to written arithmetic and paper ledgers. Instead, someone who fixes computers gets it going again, and they get back to work as soon as that's done.
Obviously they're both supposed to be proxy measures, not realistic scenarios. I was mostly joking before but I do think exams provide a pretty good proxy for ability in the subject if the teacher is decent. Interviews not so much unless the applicant is similarly prepared with foreknowledge of what they will be tested on and had some time to prepare and given recent practice.
> It’s pretty artificial in general; in “real life” you have the ability to go around online and look for sources.
I dunno how you work, but I'd be getting raised eyebrows from people watching my hit google for any question required of my role.
I mean, we're not talking about using calculators here, and we're not talking about vocational training (How do I do $FOO, in docker? In K8s? How do I write a GH runner? Basically any question that involves some million-dollar company's product).
We're talking about college stuff; you absolutely should not be allowed to look up linked lists for the first time during an exam, copy the implementation from wikipedia, port it to your language and move on.
In the real world, we want people who mostly know what to do. The real world is time-constrained (you could spend 2 hours learning to do what they thought you could do based on your diploma, but they'd be pissed to find out that you need to look up everything because that's how you coasted through college).
Exam situations are more like the real world than take-home assignments: High-stakes, high-pressure, timeboxed.
If your real world does not have high-stakes, high-pressure, timeboxed tasks, then you really haven't had much contact outside of your bubble.
To me, exams are just like democracy. The worst form of evaluation, except for all others that have been tried.
This. All the drama articles in the media are describing lazy institutions.
My tests are almost 100% in person. Project work included, you can hand something in, but I'm going over line by line and ask what you did there.
I can do this, because while my school hasn't updated the tests yet, my classes are small and I can do all of them in-person.
[dead]
So at 50%, someone who uses AI to get 100% of the homework grade will earn a D (sometimes passing) if they can get at least a 20% on your quizzes, and a C (always passing) if they get at least a 40%. Did you make your exam so difficult that students who truly didn't learn the material earn less than 20-40%? Because if it was, say, multiple choice questions with four possible answers, then you can expect them to earn at least 25% just by chance.
My quizzes are written responses, psuedocode and annotating code.
While that answers their direct question, they do bring up a good point -- how often are you handing out less than 25% scores on exams? Id imagine any professor to do that to get some severe criticism that would make even a cheater pretty livid
The last point is very interesting and might keep universities relevant.
What's interesting is that as I understand, folks are using things like Google Docs for papers, and that it's (apparently) straight forward to do analysis on a Google Doc to see, well, the life of the document. How it was typed in, how fast, what was pasted and cut back out.
My understanding is that the Google Doc is not a word processing document, it's an event recording of a word processor. So, in theory, you could just "play back" watching the document being typed in and built to "see" how it was done.
I only mention this because given the AIs, I'm sure even with a typewriter, it's more efficient to have the AI do the work, and then just "type it in" to the typewriter, which kind of invalidates the entire purpose of it in the first place.
The typing in part is inevitable. May as well have a "perfect first draft" to type it in from in the first place.
And we won't mention the old retro interfaces that let you plug in a IBM Selectric as a printer for your computer. (My favorite was a bunch of solenoids mounted above the keys -- functional, but, boy, what a hack.)
TaaS -- Typing as a service. Send us your Markdown file and receive a typed up, double spaced copy via express shipping the next day!
This would take about 1 day for some student to realize you can instruct one of the LLMs to operate the computer screen for you and have it type and fake edit a document for you. The tip would spread among the cheaters and the metric would become harder to judge by itself.
Typing as a service is a whole cottage industry on Etsy.
That's certainly one way to abstractly automate a task: Just pay someone else to do it. (This is a concept that regular people employ every day in the real world.)
Another way to automate this particular task is that some typewriters have (serial/parallel) ports to connect to a computer. It's not a daunting task at all for a student who is skilled in the art of using the bot to have one of these typewrites be the output target.
Like this: https://chatgpt.com/share/69e405db-1b44-83ea-baf3-6af41fe577...
Even Microsoft Word stores revision history inside .docx files, and that’s been used to expose plagiarism. I heard about one case where a student took an existing paper (I believe from a previous year/student) and pasted it into Word. They then edited it just enough to make it look different.
However, they didn’t remove the embedded revision history in the .docx file they submitted, so that went about as well as you can expect.
Are you sure about that? I could easily see this happen with a web document link, but for a docx file the change tracking is off by default and pretty obtrusive. Basic metadata would be fine, formatting might be quirky but that's not exactly a smoking gun...
It’s been a while since I heard about it, but IIRC the professor was a stickler for a very specific paper format, so they would distribute a .docx template file with Track Changes already enabled and require students to write their papers using that template.
I also think that when track changes was first introduced in earlier versions of MS Word, there wasn’t as much concern about privacy/telemetry as there is now, so it wasn’t made as prominently obvious.
> pasted it into Word
I'd be surprised if copy/paste carries the revision history, though. Wouldn't they have had to start with the original document (from the other student) and make their edits directly, and then submit that file?
I think they mean the revision history in the submitted .docx file showed the copied paper being inserted (as a single change), and then various changes to attempt to hide the plagiarism.
Hmm, I have some old daisy-wheel printers in the closet that I've been meaning to strip down for stepper motors, maybe I should refurb them instead :-)
In general I love the idea of turning printers into typewriters. I've been thinking about how to do it with an inkjet printer.
arms race....
oh look there is a llm trained on key loggers to spew slop at your personally predicted error rate; bonus if it identifies to USB as keyboard.
You should look up the history of the Loebner Prize [1]. There’s a shocking amount of technological development in some chatbots that went toward simulating mistakes and typing patterns to make them seem more human-like.
In some of the later Loebner competitions, when text was transmitted to the human character by character, the bot would even simulate typos followed by backspacing on screen to make it look more realistic.
https://en.wikipedia.org/wiki/Loebner_Prize
Wow it feels like the Loebner prize went away right at the dawn of the LLM. Is it correlated?
Yeah I definitely think LLMs contributed to its demise. To be honest, nobody in academic AI circles took it very seriously, because it kind of devolved into a contest over who could create the most convincing illusion of intelligence.
Participants spent more time polishing up the natural language parsing aspects in conjunction with pre‑programming elaborate backstories for their chatbot's bios among other psychological tricks. In the end, the whole competition was more impressive as a social engineering exercise, since the real goal kinda became: how can I trick people into thinking my chatbot is a human?
But reading through some of the previous competition chatbot transcripts still makes for fascinating reading.
>because it kind of devolved into a contest over who could create the most convincing illusion of intelligence.
Isn't that really what all these AI companies are doing too? It sure seems like it is.
Goodhart's Law vs the Turing Test! Can our humans accurately evaluate intelligence, or will they be fooled by fakes? Live this Sunday!
I think it would be great to be revived with a different premise.
[flagged]
Hi. I got a notification for this message in my email, so thought I'd reply. My name is Steve Worswick and I won the Loebner Prize a record 5 times with my entry Mitsuku between 2013 to the last contest in 2019. The Loebner Prize didn't take part in 2020 due to Covid restrictions and never started back up again. This along with the death of Hugh Loebner and the resulting end of his sponsorship is why the contest ended. LLMs started to become more popular in 2020 but this was pure coincidence.
I'm old enough to remember a similar controversy over whether to allow calculators in math classes. While most schools were banning them to force kids to learn how to do math without them, my school went the other way. They mandated that every student had one and then changed the assignments and tests to account for it. Gone were questions that had whole number answers that could be computed in our heads. Instead, answers were complex and the only way to know whether you'd done the question correctly was to be sure of your method. They even allowed us to write programs in TI-BASIC that we could use on tests, the only limitation was that we were not allowed to share programs with other students. I discovered that rather than trying to cram for exams, I could just write a program that would solve each class of problem we were likely to see on the exam, and by essentially teaching my calculator to pass the test, I also taught myself. It was a vastly better way for me to study. It also led to my decision to major in comp sci and my career in software. I'm forever grateful to those teachers for choosing to see the latest technology as a multiplier of student potential rather than a way students could cheat to avoid learning.
So I can't help but wonder whether schools are going about this all wrong. Rather than banning the use of AI and trying to catch students who are cheating, why aren't they creating schoolwork that requires AI? These tools are not going to cease to exist. The students they are preparing are going to live and work in a world where they exist. To my mind, you best prepare students by teaching them how to use the tools most effectively, not by teaching them how to work without the tools. Students should be learning how to prompt AI without hinting it towards a specific answer. They should be learning how to double check the answers AI gives them to ferret out hallucinations. They should be learning how to produce work that is a hundred times more complex than what us older folks had to do in school. We should be graduating students who are so much more capable than any generation before them. I think we're doing them a disservice by trying to give them the same education that was given to those from previous generations. The world they will inhabit has changed radically from the one we entered into following school.
> So I can't help but wonder whether schools are going about this all wrong. Rather than banning the use of AI and trying to catch students who are cheating, why aren't they creating schoolwork that requires AI?
Because using AI is the complete opposite of "I learned programming just to make tests easier".
By learning how to program solver, you not only learned how to program but also learned the method well enough to write it.
By pawning it off to AI to solve, you have learned nothing, not even how to prompt correctly as test questions are usually formulated well enough that AI doesn't need prompt massaging to get it.
You can use AI to get some knowledge about the problem (assuming you won't hit hallucination) but that's not what will happen when you use it for test.
And if you DO want to teach students how to use AI effectively, you can just have an AI class...
>By pawning it off to AI to solve, you have learned nothing, not even how to prompt correctly as test questions are usually formulated well enough that AI doesn't need prompt massaging to get it.
If you got AI to produce a working solution, you solved the problem. In the real world nobody who's paying you cares about the method as long as you deliver results. Students taught to solve easy problems by themselves will be at a big disadvantage in the workforce compared to students taught to solve hard problems using AI.
The part you're missing is that the evaluator already knows the answer. They're not looking that you can arrive at the correct answer, but that you know how to arrive at the correct answer. If "arriving at the correct answer" just means retrieving data from a Baysian database using a Markov chain you have only demonstrated you provide no value in the chain and should indeed get a mark of zero or get recycled.
>The part you're missing is that the evaluator already knows the answer. They're not looking that you can arrive at the correct answer, but that you know how to arrive at the correct answer.
The university evaluator is not the one paying you, the one paying you is your boss or customer. It doesn't matter how highly your university professor thinks of you, if you can't solve difficult problems as fast because your university never taught you to solve hard problems with AI, you're going to be at a competitive disadvantage in the workforce when you graduate.
I don't think the entire purpose of schools is to teach you how to answer some specific set of questions; they want to improve your knowledge and skills in various domains, and the questions are merely a roundabout way to assess that. If you can answer the questions but don't know the knowledge, you're missing the most important part.
For most students the "purpose" of studying computer science at university is to get a better job and make more money. And for the people for whom this isn't the case, they're generally smart and motivated enough to learn the extra details they're interested in by themselves.
that is a trade school.. there is nothing wrong with trade school, but it is not the same thing as a general education.
There's also no reason to learn to read and write! First graders could just point their phone at some text and have it read to them, or dictate to their phone to achieve the reverse. Why learn to swim, walk, run? Machines can do that for you too!
For now there's plenty of people who are significantly more capable than AI models. Someone who fully outscources to machines will never join that club.
You have to evaluate students on their own skills before you continue their education, because at some point AI models won't be able to help them. Anyone can use some LLM to pass the first few months of undergraduate engineering disciplines, but if you got through that and haven't learned a thing, you're completely fucked. Worse, you won't even notice the point at which AI starts to fail until you get your test results.
Once the above is not true anymore, education is pointless anyways. However for now AI can at best replace the worst performers and only in some areas.
>You have to evaluate students on their own skills before you continue their education, because at some point AI models won't be able to help them.
If at some point AI models won't be able to help them, then give them assignments that reach the point where AI alone isn't enough, so they'll only be able to solve them if they learn whatever is necessary. This is what's meant by "making assignments harder". Students who learn to solve harder problems with AI will be more competitive in the workforce than students who only learn to solve easier problems by themselves. Because AI already allows people to solve harder problems than they could unassisted, but it's a skill that needs to be learned.
As an example, with AI, it'd be a reasonable assignment to ask students to write a working C compiler from scratch. Without AI that'd be completely beyond the reach of the vast majority of students.
That's great for autodidacts, but most students will be stumped by a complicated problem if you don't slowly walk them up an incline first.
Also what do you think is an appropriate assignment for first graders where "AI is not enough"? Are we supposed to give them problems meant for engineering majors?
The things you are saying at best apply to a few select areas of education and you are hyperfocusing on them. What you are neglecting is that a lot of education focuses on teaching tool use: reading and writing is a tool, CAD software is a tool, AI is a tool, even language is a tool. For many people the best way to learn to use tools is being taught by another human being. That human being has to evaluate their progress somehow. If a first grader uses their phone to have text read to them, this tells me very little, except maybe that they can at least understand spoken language to a degree.
Using LLMs effectively, especially without essentially becoming the LLMs meat-puppet, requires a set of skills many 10th graders still struggle with. Skills like putting what you mean into words, extracting meaning from text, and thinking critically about the information you are fed.
Finally there's the matter of philosophy, ethics, and politics, which also happen to be on the curriculum in some places. Are you going to let a LLM argue for you? If you have never learned to evaluate your own beliefs and turn them into something coherent that you can communicate to others, and instead let the LLM argue on your behalf, then congratulations: you have just un-personed yourself because you refused to let others help you become an actual individual in society. You're a sack of meat hooked up to a machine. ... It's probably obvious I feel strongly about this in particular.
At the end of the day, we can at least agree that people should learn to read and write? For now?
By this logic cheating off another student or calling someone outside the room for answers should also be allowed
Why would someone hire you if all you have proven is that AI can solve the problem, not you ?
And another question: why you are even at school if all you do is put questions into AI and pass on the answers ?
> Students taught to solve easy problems by themselves will be at a big disadvantage in the workforce compared to students taught to solve hard problems using AI.
What hard problems could students solve with AI that requires the students to be especially trained? It seems you are thinking of GPT-3 style "prompt engineering". That's a thing of the past. Students can just copy the assignment into the LLM. They don't need to be taught to do that.
“You” being the operative word in You solved it. Typing a question into AI and having it solve it is not “you” solving it by any stretch.
There are a number of assumptions in what you say that don't necessarily hold.
1) That school is simply about landing a job.
2) That there is a value in students knowing how to have the AI do problems for them.
3) That follow-on effects of manually solving difficult problems is discountable compared to the direct output of the work.
I would say you're absolutely correct in that people pay for the result and they don't really care how you got there. But that's a pretty shallow rationale which overvalues the ability to be the conduit from the source of requirements to the final output and undervalues the individual ability to think for one's self when faced with the challenges of technological, geopolitical, or simply uncontrolled personal circumstances.
"The conduit", who you seem to be believe is the one with marketplace advantage, is exactly the person I would say is the most vulnerable. Not because getting the AI to produce demands is without value, but that its quickly becoming a task that doesn't need the intermediary at all. Those magicians that can prompt/agent/mcp/etc their way through to positive successes are actively being challenged by the very AI producers which our conduits people now depend on. Removing the need for intermediaries would be a great competitive advantage for any AI vendor able to achieve it. But insofar as intermediaries create output from LLMs, they'll not be very well differentiated: the common wisdom tends to be the output, lest the AI be accused of hallucination or being overly supportive. But when everyone is using AI for everything the opportunities will be in arbitraging that which is missed by common wisdom... filling in the cracks that any responsible AI would simply never venture to consider. Our conduit-person will be at a decided disadvantage because it takes real thought to know when it's best to color within the lines, and when it's best to not do so.
And that's really it. A good education is teaching you about the process of thought and becoming practiced at thinking. I would expect a better educated, thinking person to more easily adapt and make use of technology such as generative AI to solve problems more so than a person that just knows how to deal with today's prompting needs. The thinking person will be able to understand the bigger picture to better get a consistent and high quality series of results than the person just getting results as needed.
And that's really it. The output of a good education is you as a thoughtful & knowledgeable person: the output on the page is merely a means to that end. But if you focus solely on the answer on the page and the only important thing... you're really evaluating the AI, not the person that acted as intermediary.
In otherwords, if the person following your advice comes for a job, simply ask them which AIs they used in the interview and then just sign contracts with those vendors instead... you'll get better bang for your buck cutting out the middleman.
> Rather than banning the use of AI and trying to catch students who are cheating, why aren't they creating schoolwork that requires AI?
That is way to high recurring cost that many won't be able to afford. One could get a second hand calculator or even computer, and then additional resources needed was one's willingness. With mandating AI usage, we'd only increase the gap of haves and have-nots. I personally do not like the idea.
This is probably the fairest counter argument I’ve heard. One can hope that today’s AI will eventually be as cheap as a calculator, though.
I hope so too, however cheap is relative. One's ordinary morning coffee is a full day wage for someone else. If we could have decent models fitting laptops of most students, that would be point where we could possibly treat AI as we treat calculator or computer today.
Just to put things in context, https://www.bbc.com/news/articles/ce8444gex65o shares income for a good number of people now a days. (note that many of those workers are taking care of a family of 2+ members, most of the time)
I remember a TI-89 being mandatory for my AP math classes (calculus and statistics). It was utterly essential for solving problems in a reasonable amount of time. There were programs available to assist families who couldn't afford one so their children wouldn't be left behind.
I like to think we'll figure this out.
AI in it‘s current phase, definitely. However, we‘ve been seeing the transformer architecture plateauing in the last couple of years. There are still improvements, but open source models are catching up.
I feel like at this point it’s an inevitability that given enough time, capable models will be cheap enough for everyone.
If poor students have capable models but rich students have much better models that go the extra mile for a great mark and do everything in a single prompt, it would still be unfair.
For it to be fair, you would not only need good free models, but actual parity between free models and the highest subscription tier the big AI companies can offer. And I don't think that will happen in the short or mid term future.
When I was in AP classes in high school, you were required to have a TI-89 calculator. If you couldn't afford one, there were assistance programs.
You were not allowed to use a TI-92, which was the next step up. It had built-in solvers for many kinds of problems.
I'm not saying this isn't a concern, but addressing financially-based inequities in learning is not a new problem within certain bounds. There's established ways to deal with it. If we can get AI cheap enough that you can cover a year of education with $100 then we're in a good range.
That is my hope. At the same time, feels like a peak “don’t know what we don’t know” situation
A critical difference between a calculator and an LLM is that a calculator doesn't make decisions. A calculator performs the operations you type in, nothing more. An LLM does make decisions. The human operator of the LLM needs to be able to evaluate the decisions made by the LLM. That requires education and experience beforehand.
An LLM is a force multiplier only, not a replacement. It's a personal assistant to an expert. To use an LLM in a acceptable way, you still first have to learn how to do what it does yourself. I think your suggestion for people to be taught how to use LLMs is justified, but they should do so only after first being taught a no-LLM curriculum. I think this should be entirely after what the notion of an education was in pre-LLM times. Don't incorporate LLMs into our current education, instead teach use of LLMs after our current education.
continuing on personal asistant analogy, i bet even now we have ultra rich people who are not smart enough to do things themselves but are smart enough to buy (hire) smart people to work for them. And even allow them to make decisions without understanding them. But with only a guard rails : does this produce wealth for me. If yes, do what you need, i don't care :)
So this i think is applicable to AI also, pay for smarter than you AI's pit them against each other, let them supervise each other and measure the outcomes you need. Who cares how they achieve that (sound clinical and scary)
It has gone well enough for these rich people because their hired assistants were human. With humans there is some degree of responsibility, liability and a slow tempo (= time to correct) involved. LLMs are a significantly different beast.
The post yesterday about the teacher who gave students an Apple II and taught assembly was very enlightening and example of how to go forward.
Rather than banning the use of AI and trying to catch students who are cheating, why aren't they creating schoolwork that requires AI?
It's an interesting debate, but I see several reasons not to.
1. As a human you need to learn gradually, e.g. in CS you need to learn the basics of programming before going into more complex stuff. If you embrace AI from the beginning, it can let you skip the basics (why would you code a simple 200-line program if the LLM can do it?) and then you don't have the fundamentals when you reach the more complex level where human thought is needed. It's a similar problem as firing the juniors (because AI can do their work) but then who will become senior if you have no juniors.
2. If you evaluate coursework with the expectation of students using AI, those who pay the $200 subscription will have an advantage over those who pay $20, and in turn over those who use free LLMs. The only way to make it fair would be to provide all students with the best available LLM.
3. While I have heard the analogy with calculators many times, I feel that LLMs are at a different qualitative level. The calculator doesn't really replace human thought, or if it does, it's only some very specific form of it. LLMs replace human thought in a very broad way, so I think they are much more dangerous for learning.
I agree with all these points. The algorithm to multiplying being done by the calculator, is not the same because you still have to learn why you are multiplying. If you rely on ai you may not know why the ai is giving the answer. All you’ve learned is copy paste. Figuring out how someone can use ai and give the wrong answer will solve the a huge problem. But it feels quite difficult at the time being. Ai is so general it’s hard to think of how to pose the question in a way that ai can’t answer. Maybe submitting prompts and judging prompts that bypass understanding. What if the test was more how to teach the ai? Meh ai in education is filled with gotchas.
The issue I have is that people are still trying to shove AI into a pre-AI educational paradigm. And yes, if you introduce AI into a world where people are still trying to teach kids the way that they were taught in the 20th century, AI looks like a threat because it allows kids to cheat, both testing and themselves. But we have the option to stop trying to teach kids the way that they were taught in the 20th century, which I think people here are fundamentally not understanding.
It's an old book now, but Neal Stephenson's Diamond Age includes the vision that we should have for education. We literally have the tools today to build his fictional "Young Lady's Illustrated Primer." What he envisioned was not that far off from an iPad with a Claude subscription where Claude has specific goals for the conversation. It's not teachers lecturing a class, it's individualized education where an AI teaches students at their own pace using their own interests. And built into AI is the ability for precocious kids to go beyond the curriculum, either on tangents or to more advanced subjects. This is impossible in a world where a teacher is trying to shepherd dozens of students through a curriculum as a group.
In the 2010s, we got some of the way there with Khan Academy. It was genuinely new that a student could rewatch something until it clicked rather than having to digest a lecture and have any question that didn't immediately spring to mind go unanswered. AI offers the possibility to go a step beyond this. Instead of rewatching the exact same content, AI can present it to the student in multiple ways based on a student's confusion and keep explaining it until a topic clicks. It can find examples of things that a student finds interesting to show how what they're learning isn't just theoretical. If a student likes space, the AI can discuss how the trig concepts they're learning apply to the Artemis II mission. If they like sports, it could apply the same concepts to tennis. Students in literary classes could read different books according to their interests while AI ensures that they understand the same sorts of concepts while discussing them. By customizing based on the specific curiosity of the student, it can make learning far more engaging and actually fun.
To address your #2, schools should be working with Anthropic, OpenAI and Google to shape a new personalized paradigm of educating students. They should be working out deals that give access to AI to their entire district. If I were heading the Department of Education, I would go a step further and get companies to bid on a contract to put their AI in the hands of every public school child in America. A version of the AI where teachers input their curriculum into the AI and students work through it with the AI, either alone or in small groups and the AI reports back to teachers so they can intervene where they are most needed would allow school districts with staffing shortages to serve more students more efficiently and with better results.
Sometimes it feels like our current system of education is only secondarily concerned with students actually learning and the primary concern is testing students to sort them into different tiers to be absorbed into different strata of our workforce. AI does compromise this sorting process to some extent. But if we can get back to the true mission of education and think creatively to deploy AI to best educate students, we have the potential to transform education like never before. What if we don't need to test students? An AI can give an individualized assessment of how well a student has grasped what they're supposed to be learning based on weeks of individualized work. It's as if we can give every student their own private tutor who will report back to the teacher on the student's actual progress. When you have that, stress-inducing exams are a ridiculous substitute.
I've been pretty shocked at how closed-minded the responses to my comment have been. We're supposed to be a community that envisions radically better futures that can be built with technology. And here we have a revolutionary new technology that upends a staid and increasingly problematic part of our society and the majority of the responses are geared towards explaining why that staid and problematic institution should be maintained unchanged. AI is fundamentally a danger to our current education model, but that model can change radically for the better. And I would've hoped that more people here would have recognized that.
Have you tried to learn a new topic with ai? Not like, learn the answer to a fleeting question. An actual large topic one might take even a short course on.
I say because most of the experiments I have seen in this space have failed. Chatgpt education was quietly removed not that long ago. Khan academy recently said that their Khanmigo AI tutor was facing challenges because students don't want to use it. Its a long-standing observation in the field of education that the miracle of the internet, computers being ubiqutous, phones etc hasn't clearly resulted in improvements to education (theres some minor evidence for or against it, but no blindingly obvious effect)
I worked at one point in ed-tech and the longer I was there the more I realised that nobody wanted this. Students only used it if they were made to, teachers only did it if admins wanted, admins only did it if they were sold on it, and the sales people seemed to be the only people who actually thought it was helping anyone
People often seem to think that teaching is the process of neutrally presenting facts to a learner. That the better and more clear the facts are presented, the better. But any book can do that. The entire business of education, as it were, is the management of motivation in students. Exams function just as much as a tool in this as the institution of coming to a classroom at all, rather than sitting at home remote learning. You need a clear mind, distraction free - just bored enough to find the content acceptably appealing (content cannot be made more interesting).
The best things for education seem to be:
- getting enough sleep
- getting a good diet
- good exercise
- being around other people you respect who are also learning the content
- being around other people you respect who have already learned the content, and who you want to emulate
- having someone who visibly cares if you learn the content or not, who expresses and reinforces that they expect you to learn it. (both "you need to focus, stop messing around" but also "you are capable of this, its hard but I know you personally can do it")
- having a good reason to want to learn the content
- having time pressure to learn the content
note that literally none of this regards the actual presentation of the content. Books have existed for centuries. A well motivated learner with all of the above will find the content. That is in no way the problem
An argument might be "but why, given all of this, are teachers actually teaching, why arent they standing up and lecturing soley about the importance of learning the content" and to some extent thats fair (although lecturing does satisfy some of the above if you think about it), but they ALSO tend to assign reading, link to further resources etc. Probably half or more of teachers speaking time is given over to procedural stuff working towards the above goals. Explaining the process of upcoming exams, worksheets, homeworks that need to be done (time pressure), demonstrating their love and knowledge of the subject (people you respect who know the content), building rapport with students (people you respect), holding people to account/motivating, explaining why the content is important, and trying to build good habits in their students (organising study groups, project work, dealing with problems and creating distraction free conditions)
AI are a poor facsimile of the above. Though they may try to replicate it, their inherent lack of physicality and humanity makes much of the power lose its effect. I dont care if a chatbot is disappointed in me. I'm not inspired by a chatbot that claims enthusiasm about a subject. I don't care if a chatbot tells me I absolutely must do this by next week else ill be left behind (behind who?).
I think the controversial part about AI is far above the "it is a new technology so use it".
When the scientific calculator was invented, people could easily know what went into its production. As in what circuitry appears in them. You knew that if you bought it, it is yours. Want to program it? Grab a book and do this. The whole package would be a fixed price. You are in control. With AI? You are not at all in control. You rely on a big tech giant (or just like 4 useful ones) who is riding what people controversially still call an economic disaster. You are relying on a technology that is designed to very likely bait-and-switch you. As soon as you get too comfortable with AI, the big tech companies can just bump the prices up and you will not be able to say no. You rely on a technology that you do not control.
The comparison of AI to a calculator or any other technological advancement for students is apples and oranges for that reason.
Imagine giving a student a personal AI datacenter to carry with them. This may be more of a fair comparison.
PS Training students on using AI, especially for free, is setting them up for reliance on the big tech companies and the subscription model.
Your comment is framed like "giving a student a personal AI datacenter to carry with them" is unrealistic, but in fact it is easy for anyone with access to $1000-$2000 worth of compute to download and operate exactly that for free, with performance perhaps a year behind the state of the art.
> but in fact it is easy for anyone with access to $1000-$2000 worth of compute
Even if we assume that to be true, you severely underestimate how many people that condition excludes.
There are simpler LLMs that run on much cheaper devices and are still helpful for baseline tasks. Of course they are prone to hallucinating once they reach the limits of their world knowledge, but this also changes their effectiveness in an educational context: they can help you polish a paper (much of their reliable knowledge is about language, syntax and style/pragmatics of the input texts), but you still have to plan the writing on your own.
Maybe teaching students to take their whatever devices to run AI is the way, sure. All I tried to say is if we're teaching students to think independently, we should teach them independent tools.
It doesn't exclude people who attend high schools and colleges that have a computer lab.
That however requires significant investments - either each computer gets a powerful GPU for local inference (which cost a fortune) or the school gets a rack worth of compute. Most schools however even struggle to get their children fed.
Another issue is that it forces kids to stay in school for longer to do their homework, which can be a serious problem in rural areas where public transport is limited, so parents are forced to fetch their kids from school which may not be compatible with working hours.
Replace "using AI" with "asking your parents". From a student's perspective, their parents are probably an expert in anything, but sometimes might make things up and they won't be any wiser to notice, because they don't yet have the basic knowledge to know what to double check for. Just like LLMs.
Why doesn't the essay class allow us asking your parents to write it for them? The art class, why not ask your parents to paint something for you? Geography, why not let ask your parents during a test?
The difference is you learned useful maths teaching your calculator. At the moment a teacher can't tell if you even read the LLM output. Even if all future literature is writen with an LLM it is highly likely the are skills you need to learn to become a best selling author.
So, scientific calculators
- made tasks easy that were a necessary prerequisite for advanced math (basic arithmetic), but not what the lesson was supposed to be about
- could in theory also let students skip over what they were supposed to be learning (applying the correct operations in the correct order to solve a problem) but doing so would require programming or getting a program from someone else, which the teachers probably figured was a high-enough hurdle to accept the risk
Hence, scientific calculators helped teachers by removing unnecessary friction.
Meanwhile, current LLMs
- will happily attempt to do the student's entire homework for them
- cannot reliably be restricted in functionality to leave the part the students are supposed to do themselves to the student
Hence, LLMs undermine teachers by removing necessary effort.
Sure, in theory LLMs could enable even more focused lessons by removing even bigger unnecessary frictions (e.g. in history class, have a LLM scour a large collection of primary sources to exhaustively list passages mentioning a certain topic), but students cannot be trusted to use them this way.
Hence, teachers are trying to use all kinds of tricks to ensure that what they wanted to teach actually passed through the student's brain at some point.
> - cannot reliably be restricted in functionality to leave the part the students are supposed to do themselves to the student
Small local LLMs are essentially that. If an LLM can tell you to eat rocks as a tasty snack or use glue to make the cheese stick to your pizza, imagine what it says when you ask it to analyze/explain complex academic subjects, or solve fiddly problems. But it will still reliably help you polish your language, like a subject-specific dictionary/thesaurus.
Small LLMs are certainly restricted, but they're not reliably restricted in a way that aligns with educational goals. Even Qwen3.5-0.8B, the smallest, dumbest LLM I have lying around, can solve quadratic equations. In math class, learning about quadratic equations for the first time, such helpfulness defeats the purpose. In physics, learning about ballistic trajectories, it might be acceptable, as long as the students themselves come up with the correct equation to model the problem, and do the reasoning to identify which of the two theoretical solutions corresponds to the physical setup.
A single model that can do many things, some less reliably than others, independently of what the requirements for the lesson are, is not a good tool to give students. You would need something that can do ancillary tasks perfectly, but won't do the part student are supposed to practice doing themselves. And what exactly that is changes with every exercise.
If you want to give students access to thesaurus functionality and nothing else, you're better off with a thesaurus.
Homework is waste of time for everyone involved tho
Because AI is different from calculators. People stop thinking and just excel at verbose and incomplete rationalizations of every evil, just like you do in your reply.
Yeah... We had those bulky TI Voyage 200 graphical calculators in school [1]. They could do everything the teachers could throw at us up to the point of having all but a few formulas build in.
I would say that definitely shaped me in a way where I rarely bother with the underlying details and tend to focus on how high level abstractions interact. [2]
[1] German "Mathe-LK", we could chose specializing in two things, for me it was math and computer science, the later being quite novel back in 2003. [2] I _do_ tend to specialize in things, but e.g. for LLMs or GLMMs, while I do have the capability to understand the technical details, I just don't bother.
At this stage, I can no longer take comparing LLMs to calculators as a good faith argument. That’s a talking point, a framing device, one whose flaws have been explained ad nauseam (as exemplified by the sibling replies), and I’m left questioning either the reasoning abilities or the honesty of those still making it.
They. Are not. The same.
Have you ever known people to commit suicide, kill, or give themselves rare diseases because of their calculators? How about people dating their calculator and going batshit for a software update?
Not to mention learning to do on your own is a useful skill to teach you to think, and an essential skill to (as you suggest) verify answers. People not understanding how things work is exactly why they take bullshit output from an LLM as gospel.
I also note that such arguments tend to be profoundly selfish and self-centred. Your anecdote happened to have an outcome you enjoyed and benefitted from, but I bet that wasn’t the reality for all your colleagues. Just like you are glad for the calculators in your class, some other student may be glad for the lack of them in theirs and it may be the reason they got into their field of study.
We weren't allowed to use calculators until aged 16 and I'm glad because I learned to do mental arithmetic. Many people I know, even in STEM, aren't able to do that, or lack a feel for numbers.
It was good for you but you don’t address reality of life.
Here’s one possible scenario: After graduation, you (or somebody else) shares the program with friend, with a promise to not to share further. Soon enough, it’s on everybody’s calculator. What did real educational thing for you, is just cheat where one needs to press the right buttons and get the right answer. This completely destroys the educational purpose, but significant amount of people just don’t care and want to get a pass.
Yes, there always is a counter weapon by teachers: for example, to point to random line and ask to explain and whatever, but this is not (always) scalable.
I’ve seen this in reality in college, when there was a cs/database course final project implementation, written in Delphi (very popular at a time in xussr), that was passed from year to year, that the professors and ta were so fed up, that I got almost auto pass because I wrote mine in C++…
——
To summarize - the overinreasing amount of pure slop is seen everywhere. Regular multi-thousand line prs where author didn’t even bother to look into code, written by ai. Just prompt -> commit, push, or. Nobody wants to deal with that
Same is happening here - u it’s not to punish people who use tool in proper context, it’s to filter out people who just don’t give a fuck.
The calculator analogy comes up very often, and it’s a good one because it also illustrates where AI diverges.
The other analogy is taking a forklift to the gym. Sure you lift weights, but you don’t really do any exercise to develop your own muscles.
AI automates a significant chunk of the exercises. So you are left with people who didn’t build any mental muscles.
This would be bad enough, but it’s worse because AI severely benefits experts who have build mental reflexes/taste and can judge / verify output with minimum information.
Interesting thought! I wonder what kind of assignments, etc. would require AI that would still allow assessing how the student has absorbed the knowledge. Maybe the assignment would require sending in the prompts and the whole conversation how the problem was approached and solved?
With the calculator analogy, I think that the calculator automates executing certain algorithms (like multiplication, etc.), but using AI takes away some (most?) of the thinking.
Bad idea. This is pretty much endgame territory you are talking about.
You would give the brains of the younger generation to American tech oligarchy, a class of people openly hostile to the principles of the democratic rule of law. If you want to see the damage actors like Fox News et alii alone can do, just take a look around in the US. Now imagine them taking over the parenting and teaching role; you wouldn't need gerrymandering if you can control people's beliefs.
There’s one significant difference between: typewriters and calculators are a one-time-paymeny device. LLMs are subscription based baked by billionaire companies. That alone leaves LLMs in a bad enough place
You have a point, but it is very ironic that this sounds a lot like the old argument "you must still learn to compute by hand because you won't have a calculator with you all the time".
Just like computers in schools nowadays, I will concede that part of the education should be about learning how to use AI - but they should also learn how to do research and accomplish things without the help of AI.
>why aren't they creating schoolwork that requires AI?
Because this makes a subscription a requirement for education, and thus advances the grift that is subscriptions, rent-seeking and dependence on a service. This isn't something we should ingrain into our children from an early age.
Calculators were buy once, use forever. Subscriptions to slop generators are a long term dependency and I want my children to not be exposed to that until they can decide for themselves.
In the Uk we have two exam papers that make up the age 16 mathematics exam. In one you use a calculator and one your don’t.
You need to be able to do both things. We don’t need to make it a choice.
> So I can't help but wonder whether schools are going about this all wrong. Rather than banning the use of AI and trying to catch students who are cheating, why aren't they creating schoolwork that requires AI?
Back in college, my assembly class was in MIPS (incidentally taught by Professor Larus of SPIM fame). I remember slogging through writing the assembly to compute factorial and saving registers and then dealing the frame pointer and the stack pointer.
One of the other students had access to a DecStation, wrote the program in C, and ran gcc -S to get the MIPS assembly from it. However, the compiler realized the for loop (and tail call) optimizations and instead of making it a recursive function (and help us practice and understand $fp) write it with a jump instruction instead.
Aside from getting a 0 on that homework, they struggled with the next assignment that presupposed understanding of how write function calls.
---
You could argue that learning C makes needing to learn assembly irrelevant (and MIPS is even less relevant today than back in the 90s). But for learning in school, it's not about the assignment but rather the journey that one takes to get to that assignment and learn from it.
Being able to check the answers that are provided to someone requires the understanding of what goes before and beneath the answer itself.
Writing the assignment in C when you're learning the before and beneath of "this is how assembly works" means that when you later take the complier class you won't be able to debug if your code generation is incorrect.
Working with an AI as the primary tool for learning problem solving keeps the person at the higher level. There is some foundational level that a person needs to learn without relying on an AI to do it for them.
The AI and other abstractions of the underlying problem do allow us to work with more complex problems. Would you trust a bridge built by an engineer who built it with AI and didn't understand the underlying math and physics themselves?
This is especially at issue in college where some students are taking classes to get the requirement out of the way, some are taking classes for that directly - an accountant is taking the math class to do math with the numbers, while the engineer is taking the math as a prerequisite for physics which is a prerequisite for a material science class which is a perquisite for a soil mechanics class.
If you don't understand the various foundational levels without using AI then trying to identify where the AI (or any other tool) got it wrong isn't something that you're necessarily able to do.
There's a big difference between learning to program a calculator, which is deterministic, and "learning" to prompt an LLM-based AI service that is tuned deliberately to be non-deterministic and that changes every few months. I would prefer my kids learn things like thinking critically and communicating and logic, not spend their time "chatting" with an unpredictable oracle.
It's not a bad point, but the "calculator" metaphor is not the best one. First, schools don't create schoolwork. Teachers in different disciplines do. Second, calculators aren't in pervasive use across disciplines; AI products are. Third, schools provide these AI products for free - teachers don't. So a student in a psych methods class uses a free tier to create lit reviews, or whatever. The asymmetry there creates a problem for most assessments of skill, because students want expediency, and teachers assess skills and memory.
Finally, nobody seems to know what teachers are actually discussing. To your assertion:
>So I can't help but wonder whether schools are going about this all wrong. Rather than banning the use of AI and trying to catch students who are cheating, why aren't they creating schoolwork that requires AI?
Many teachers do create schoolwork that requires AI. Many teachers ban it. Everybody's trying to work out policy (to the detriment of other policy discussions, particularly the new ADA landscape). Many ED departments are captured by AI vendors - AI is a normal technological competency for ED majors at different levels. It's not that the discussion is not "are students going to work with AI?" The discussion is "how do we teach?" which is what the discussion always is.
But policy is a part of that. Admin will have guidance and policy statements, and each instructor will as well. Students, who get thrown off balance if they have two teachers with different nav bars in the CMS, want clarity of policy: can I use it? is a different question than should I use it? But "should I use it?" is the much more relevant question for instructors. The instructors passed through the 90s/00s/10s blissfully unaware of anything that was happening in these fields.
>These tools are not going to cease to exist.
Which tools, precisely? Because I'd assert your 200-dollar-or-whatever tier that runs out of tokens on Monday does not functionally exist for most students. I don't know what happens at MIT but Penn State satellite campus students aren't whipping up agentic solutions to "I have a summer online course with discussion boards." They're just plugging that shit into whatever chatbot they have on their phone. Honestly, most online courses aren't even worth that effort, but: different discussion.
The only reason there is pervasive student use is because someone made it free. The CoPilot window that comes with the basic tier of 365 and all the other in-app copilots are what my students have; the Google Docs stuff exists, and the Grammarly stuff exists, and the best way to "ban" it (which is just fine as an approach) is to make it even slightly more expensive. If someone does that, yeah, I think some of these products might cease to exist.
>I think we're doing them a disservice by trying to give them the same education that was given to those from previous generations
Students graduate into a world where all kinds of stuff is going on and all sorts of ideological forces shape what they encounter in schools. "AI" more a dominant element in that force than the calculator was because calculators didn't have hundreds of billions of dollars of investments, and I'm not sure I ever knew what the prevailing political project of Texas Instruments was. Maybe TI's CEO had a manifesto; tbf, I do remember IBM's corporate culture being strong enough to drive cultural change. But it would be naive to not recognize that these products are made by people antagonistic to the idea of education as a compulsory and public good.
> The only reason there is pervasive student use is because someone made it free. The CoPilot window that comes with the basic tier of 365 and all the other in-app copilots are what my students have; the Google Docs stuff exists, and the Grammarly stuff exists, and the best way to "ban" it (which is just fine as an approach) is to make it even slightly more expensive. If someone does that, yeah, I think some of these products might cease to exist.
So the solution is raise cost, thus making the richies have access but the poors not to?
I guess thats one way to academically bin students; just put yet another financial gatekeeping to the academic process?
Bit silly to read that as a "solution." Financial gatekeeping? Unis don't set costs for AI products.
>thus making the richies have access but the poors not to?
I don't know that you need "thus" to describe the status quo. My school has tiny little Dells that are ten years old inside our Prometheus units. You think we're gonna upgrade each unit in the school with costs as they are, or do you think MS is just going to suddenly "provide" a cloud computing solution?
Imagine military instructors would say that it's important to focus on cavalry and bayonette use just because they can't figure out how to adapt their curriculum to the new reality.
World war I was very much that.
Why are people promoting the idea that exams are not written or given in person anymore? I graduated relatively recently and maybe had 1 take home exam during my entire education. Every other exam was proctored in person and written. The professor who made the take home exam also made it much more difficult than a normal exam so I would not really say it was easier than a normal in person test.
Things have changed drastically since COVID-19, at least in the US. Tons of schools and universities shifted to online systems, and never abandoned the systems they built up when it was time to go back to school.
I graduated in 2020, so I've only gotten to see the changes secondhand through friends and family who are teachers, and through my sibling who graduated a few years after me. But the difference is staggering.
I loved take home exams because they allowed me to study before hand but not have the insane pressure and condensed studying required for exams in the classroom. Even though they were normally much harder and longer, I liked them. I felt I learned much more through them because I could take the time to understand concepts I had missed without feeling the time pressure of in-person exams.
It's a shame that humans find a way to cheat ourselves out of things that benefit us by over "optimizing" the wrong things.
Exams in classroom with all the time pressure is also an important part of education. May be they should be low percentage of grade to prevent too much stress but it's am important learning experience
I disagree. Take home exams represent how work and progress occurs in the "real" world. There's nothing in the post education world that resembles in-person exams.
Maybe the medical profession is a counter example.
> There's nothing in the post education world that resembles in-person exams.
I’d argue that dealing with any high criticality operational incident is like an in person exam (maybe even the most difficult kind, the open book one) if you are the one responsible for fixing it. Everyone is looking at you, you have time pressure to solve it ASAP and you can’t afford the time to dig through all the docs on the spot. So there’s at least some similarity with some real life situations.
I can understand that argument, but I don't think in-class exams are specifically good at that. I think for those situations, it requires more specific training regarding the domain, which is why I gave the medical profession as a counterexample.
I'd like to see some data on this. My general-ed recall is minimal, and in programming before school, I certainly learned a ton more by coding than by testing. That's my perception of my time in school, as well.
Take home exams were very common when I was in school, which was before you could get answers on the internet. After internet answer and cheating sites came along, a professor would have to either not care and let cheating run rampant, or struggle to constantly make unique new kinds of take home questions somehow. AI has basically killed that option too.
Did you by any chance graduate before the COVID-19 pandemic?
on the flip side, surely people know that people cheat all the time on "proctored in person and written" exams
A lot of people in this thread are talking about how they did in-person exams, handwritten problem sets in class, etc. This kind of thing is more challenging in the humanities, where the research paper is kind of our bread and butter. A lot of us have since turned to different kinds of assignments, but I am not ready to forego research papers in favor of blue book exams. I think there's some value in having to develop and sustain an argument in conversation with some body of literature (scholarly or otherwise), and that is not easily to replicate with in-person writing, at least at the undergrad level. (Doctoral candidates do this kind of thing all the time in qualifying exams, but that's after years of graduate school and fresh off doing nothing but reading 100+ books over the course of a few months.)
My undergrad was in humanities - we had multiple essays to submit throughout the year but they only counted for ~10% of the final grade, with the rest through an in person exam paper.
It was just expected that you had a grasp of the literature enough that you could argue off-the-cuff in the exam setting, and then you were given leeway if you didn't have exact Harvard style notations to exact date/titles of referenced material.
It's easy to imagine a paper/exam combo.
Student turn-in a mid-term paper. A professor chooses a certain number of points from the mid-term paper and asks for explanations of these in long-hand. Pulling questions this way doesn't seem like it would take more time than a thorough reading of the text.
Oh, but paper reading has been delegated to a drudge you wouldn't trust with pulling question, oh how inconvenient. Which is to say the problems AI introduces to education are strongly related to much of work already being made mindless before AI appeared.
Doctoral candidates do this kind of thing all the time in qualifying exams, but that's after years of graduate school and fresh off doing nothing but reading 100+ books over the course of a few months.
No, High school students can do this. Well, they get impelled to do this. They can't do this now but that's a testament to current education.
I'm not sure of if typewriters help for assignments. People will just rewrite what the AI wrote
That's the first thing that happens.
The next thing that happens is a befuddled "Ask HN: Why Do All The Daisy Wheel Printers On Ebay Suddenly Cost Thousands Of Dollars?".
https://en.wikipedia.org/wiki/Daisy_wheel_printing
You can sometimes locally win an arms race by doing something really exotic that isn't worth the work to defeat, but this is definitely not a strategy that works if everyone adopts it.
In one of my classes the approach was the opposite, I’m expected to do Ph.D level work as an undergrad and am expected to use AI.
In a different one she just said so long as you say AI was used you’re fine to use it.
In the rest of them AI is considered cheating.
To say we have discrepancies in the rules in an understatement. No one seems to have the exact answer on how to do it. I personally feel like expecting Ph.D level work is the best method as of now, I’ve learned more by using AI to do things about my head than hard core studying for a semester.
If it’s any consolation, this problem of discrepancies in rules is very common at universities now.
I teach at two universities in Japan and occasionally give lectures on AI issues at others, and the consensus I get from the faculty and students I talk with is that there is no consensus about what to do about AI in higher education.
Education in many subjects has been based around students producing some kind of complex output: a written paper, a computer program, a business plan, a musical composition. This has been a good method because, when done well, students could learn and retain more from the process of creating such output than they would from, say, studying for and taking in-class tests. Also, the product often mirrored what the students would be doing in their future lives, so they were learning useful skills as well.
AI throws a huge spanner into that product-based pedagogy, because it allows students to short-cut the creation process and thus learn little or nothing. Also, it is no longer clear how valuable some of those product-creation skills (writing, programming, planning) will be in the years ahead.
And while the fundamental assumptions behind some widely used teaching methods are being overthrown, many educators, students, and administrators remain attached to the traditional ways. That’s not surprising, as AI is so new and advancing so rapidly that it’s very difficult to say with any confidence how education needs to change. But, in my opinion at least, it does need to change at a very fundamental level. That change won’t be easy.
It's not inherently contradictory, just like using a calculator could be considered cheating depending on the context. If you're just learning basic arithmetic, a calculator is cheating since it shortcuts the path to learning. OTOH in calculus, a calculator is necessary. You still have to have a deep understanding of the concepts and functions to succeed.
It's still a new tech so I'm not surprised a lot of teachers have different takes on it. But when it comes to education, I feel like different policies are reasonable. In some cases it's more likely to shortcut learning, and in other cases it's more likely to encourage learning. It's not entirely one or the other.
A better example might be physics and math classes. I was learned derivatives and integrals at the same time in those two classes, but the math one required we learn how it all works (using limits to understand why the derivative rules work, without using calculators, for example), while in physics we just memorized the rules and were expected to use the calculator.
I always thought they should teach calculus first.
Why do you need a calculator for calculus?!?
Gotta add up all the curves
Exactly, AI is the next calculator. Right now the consensus is that it just does the work for you, in my opinion that says more about us not having the right questions than actual laziness. In a world where the only questions are basic arithmetic, calculators do all the work for you. My opinion is that the future what used to be done by academics will be done by high schoolers and new academics will be producing work at a rate no one could’ve ever predicted.
For example, the professor who’s leading me in this project had a fellowship at a certain university in England and said he exclusively coded using claude code for a month straight, their purpose was to solve a vaccine for a specific disease and by using AI tools such as claude code they’re several months ahead of schedule.
[flagged]
I'm really not seeing how you can do PhD level work as an undergrad. You wouldn't have the foundational knowledge necessary to do PhD level work, and you have no idea how much of what you're learning is accurate.
Without going into too much detail, when I said “Ph.D level” I’m meaning active research that adds a meaningful contribution to a field. I’ll probably be posting on here in a couple months about it but I’ve been doing thousands of tests with beefy GPUs on a certain theory we have about small 9b LLMs under certain external constraints.
Am I saying I’m as knowledgeable or capable as a Ph.D right now? Absolutely not. There’s just not really a terminology that correctly describes accelerated learning and iteration by use of AI since the technology is so new. I can’t speak for others but as someone who’s a senior in my physics degree, I’ve been actually learning faster by using AI. It’s either a mental crutch or mental accelerator. The difference is in if you want it to completely do work for you or if you try to learn and follow along.
It’s a very under explored and new area right now, how higher learning is effected by using AI as a tool instead of as a cheating device, but historically, new tools like the calculator or computer have done a lot to accelerate learning once new rules are in place.
For what it is worth, no graduate student would say they are doing 'Ph.D level' research. It is called 'graduate level research' or just, you know, 'research'
Sounds like a fun project, I wish you the best. I ran a similar program (independent study that encouraged freshman/sophomore undergraduates to explore using microprocessors, at the time the EE curriculum was completely focused on analog circuit theory and ended at boolean logic) and it went well enough that it eventually became part of the official undergraduate curriculum.
It's not terribly uncommon for an undergrad to claim they're doing "PhD level work".
Undergrad research is pretty common and it's not all that hard to get your name on a paper as an undergrad. A lot of undergrads think that doing work that gets your name on a paper, equates to PhD level work.
Now at least you're an adult already. Imagine what mixed messages schoolchild's are receiving from their teachers...
> I’m expected to do Ph.D level work as an undergrad and am expected to use AI.
Nice idea. What class and what work are you doing then?
For that specific one, it’s more of an independent project analyzing complex systems for 6 credits, I’m gonna be expected to submit a paper to arXiv on the subject with the professor as a co-author (fingers crossed). He said I can use claude code or any AI. I’m required to do X amount of hours per week and then submit a thorough report after about 2 months.
>I’ve learned more by using AI to do things about my head than hard core studying for a semester.
How do you know you actually learned, instead of being fed slop by the AI that isn't true at all? If you didn't study, then I doubt you'll really know if the AI is lying to you or not. I have to wonder if your teacher will too, sounds like they have kind of checked-out from actually teaching.
Reading all these comments, I feel like US universities are a joke.
I had to do all the exams in person. 100% of the grade was decided at the exam. Millions of people graduated this way and they are fine. No students were harmed in the process.
I agree fully. Not sure what they are on about, with "no labs??" in the replies.
You still do all the same things, and they are graded, but this doesn't affect your final grade. Instead, you need to pass a threshold to enter the exam, which is then graded.
The US isn't so amazing at this, it simply can be done better. Recognizing where you can improve and from whom you can learn is a great first step to ACTUAL improvement.
No projects, no labs, no teamwork, no papers?
What a narrow set of skills to send into your economy.
Just because those things don't contribute to your final grade doesn't mean you don't do them.
At Oxbridge, for CS we still had lab work. We still had problem sets assigned for CS and for math which were graded. We had one large CS group project in, I want to say, our second year. Humanities students were still assigned essays. It's just that none of this stuff contributed to your final degree classification which was based entirely on your exams (although if you didn't do your CS practicals you wouldn't be allowed to pass).
Obviously Oxbridge isn't exactly representative but certainly my experience showed me that the American style is not the only way of making education work.
oh, that's why US has the only working economy in the world ! :)
joking, of course everyone does 'projects, labs, teamwork and papers'. It's just not the main focus of the grading process.
Given the way things are going, not knowing how to use AI will be like coming out not knowing about revision control
Isn't the selling point of AI that it does it for you? What's to learn?
If the AI does it for you, you need to still learn what to do.
What is the "it" that AI does for you?
This is assuming you know how to get good work out of AI in the first place. But even that is turning out to be a skill in and of itself.
"It does X for you" is the point of many technologies. You still require knowledge to work around it.
Context helps immensely, for example. Think of what you can do that someone outside tech can't.
The “it does X for you” aspect of technology is not completely without its downsides, for various values of X.
For example, take “X” to be “walking”. Do we have the technology that allows us to pretty much never have to walk? Sure. As far as I am aware, though, we do not generally favour a lifestyle of being bound to a mobility aid by choice, and in fact we have found that not walking when able in the long run creates substantial well-being issues for a human. (Now, we have found ways to alleviate some of those issues for those who aren’t able, but clearly it is not sufficient because we still walk.)
The problem is exacerbated immensely as the value of X approaches something as fundamental to one’s humanity as “thinking”.
> "It does X for you" is the point of many technologies. You still require knowledge to work around it.
When running water replaced the need to pump water out of the ground yourself, were people urged to "learn faucets"? You kind of just need to twist a knob and water comes out, right?
Maybe there was an intermediary stage where running water was slightly more complicated and there were more steps to learn, but devoting time to learning those steps would have been a waste of time, since the end goal of the system was for it to function without much input.
[flagged]
There’s a big difference between using AI well and using it badly. Judgement is what needs to be learnt
Weaponized FOMO to drive everyone to use AI isn't really a great idea
I think you’re missing the /s.
Not at all. It’s a tool. It can be used well and it can be used badly, the difference often being others things thought in a CS curriculum. But a well trained engineer using the tool will be more productive than that engineer doing everything by hand, so leaving that tool out of the curriculum is doing a disservice to the students
So you didn't have to do any course work? No collaboration? No labs? I'm not aware of any University that doesn't have coursework outside of online diploma mills.
In my undergrad, coursework did not count towards the grade for the module. But you earned the right to sit for the final exam by passing the courswork.
We had a token portion for the grade from coursework. Eg 10%. Its enough to force you to do it, but not enough to influence the grade too much.
for people with medium to high anxiety problems and/or with ADHD, exams are an absolute nightmare that is not representative at all of workplace performance or ability to do good work.
Did you never have to write a research paper?
A million foreign students are studying in US universities. Millions applied. What a joke education system indeed!
When I was in college, your grade fully depended on the oral exam/debate with the professor. Everything else was but the entry ticket.
Not sure anyone even attempted to cheat in that scenario. And the conversations were usually great, although very stressful for us cramming types
This sounds extremely susceptible to unconscious bias, or even just straightforward discrimination.
It does! That’s why you can ask to be evaluated by a commission of professors.
If you don’t pass after 3 tries, commission is mandatory.
You also have a paper trail of written exams and midterms to back you up. If you keep getting good grades and failing the oral, people will find that obviously suspicious.
Honestly the only times I had any trouble in the orals were the exams where I baaaaarely passed the written. Usually oral feels like the chill easy part compared to written because you can have a back-n-forth with the professor.
> It does! That’s why you can ask to be evaluated by a commission of professors.
Still concerning from a statistical/psych fairness aspect.
There's a famous example of the Boston Symphony trying to fairly judge unseen applicants in 1952, and their results kept getting gender-skewed until they adjusted for the fact judges were reacting to the sound of shoes (e.g. high heels) when the candidate moved around behind the divider.
> That’s why you can ask to be evaluated by a commission of professors.
Ah yes, the classic "if you think the system is abusing you, you shall out yourself to the system that's abusing you if you want any chance of recourse." Because a tribunal run by the people you're lodging a complaint against can't possibly be biased.
Moreso than a job interview?
More systematic than a job interview.
If you don't get one job you should have - there are others - it's unfortunate but not life altering.
If 3 years into your marine biology program a professor who always teaches a mandatory course fails you because you're a woman who wears non traditional dress - you're not graduating and now there are no jobs. (And this is an example that actually happened to someone I know - not in a western country)
I think the oral exam is probably a great way to ascertain a student's ability, but let's be real, undergraduate class sizes numbers in the hundreds, for almost every first year class. I don't think it's possible to administer that. I think I would have loved to see that in my later years at university though, we still just did things by written exam + course work.
> let's be real, undergraduate class sizes numbers in the hundreds, for almost every first year class. I don't think it's possible to administer that
Our first year class was about 250 people. It was fine.
By the 4th year, class sizes were a much more manageable 30 to 50.
You get maybe 10 to 15 minutes with the professor (usually more in later years), they ask 3 questions with some followup. That’s 1 work week for the professor. And less than half the students even make it that far for every exam season (3 per school year) so you’re looking at something like 3 days of work. It’s fine.
I had a typewriter growing up and I remember thinking it was the coolest thing. I was amazed by it and tried writing several stories. Eventually my dad bought me a crappy old computer that was only really good for writing, and that was cool too. I loved that thing. It was small too, with an integrated monitor and keyboard, so it didn't take over the whole desk where I still used pencil and paper often
Imagine being able to do some writing without notifications going off every few seconds, and where you're not always one click away from a search engine and some website scientifically designed to drag your attention down a rabbit hole and keep it there
There's an entire industry of "distraction free writing devices" based mostly on that nostalgia/yearning (not to say that it isn't effective, but the effectiveness is not actually being measured :-)
I have an old MacBook Air I flashed with writerdeckOS [0]. Feels like a digital typewriter.
[0]: https://writerdeckos.com/
For programming, the way I would build a curriculum is to force students to actually learn how to program and code first. This is simple by requiring them to write code inside the classroom by hand for all exams.
I would make this the focus for 90% of the first 2 years of their degree.
I would then have them spend 75% of their last 2 years learning how to use and program with AI. Aside from knowing how things actually work, there's no more important skill now than mastering AI.
First part is basically my degree. You can cheat as much you want, but when you have to write algorithms by hand on paper in the exam it doesn’t matter - you will have to learn. Then depending what you want to do for masters it’s basically machine learning, which mostly mathematics.
I don’t know why you’d need to teach anyone to code with LMM’s though. If you are a CS major (or any reasonably intelligent person) and can’t figure it out on your own, well, workd needs bakers and carpenters still you know?
What I am learning is mathematics, which what all CS including AI and machine learning, really is. ”Mastering AI” for me means building your own models and AI applications and undestanding linear algebra, multivariable calculus and the propability theory behind it all.
One consequence of LLM fraud at scale making remote/online tests & document submission worthless is it might act as a giant revitalizing boost for the bricks-and-mortars school systems. Suddenly having real teachers and students in room together has value again, for credibility and authenticity alone.
LLMs are also making having a public repo code portfolio be much more worthless as a sign of legitimacy
I can't compose at a typewriter the way I do with a word processor. I would have to write it out by hand first. If the typing is just transcription, I could just as well be copying from an AI doc.
If you're doing it in class anyway, and providing typewriters, you might as well provide a locked down Chromebook. Cheaper and better for composition.
Writing an extemporaneous essay from start to finish in one draft on the spot is a skill that students were expected to develop in my educational experience in the US less than 15 years ago or so. If that's vanished, I'd be glad to see it come back.
> Writing an extemporaneous essay from start to finish in one draft
Every English class I had in high school focused so strongly on how important revision was, or at a minimum having an outline to work from. While AP tests expected us to dash off essays by hand in a single go I empathize with OP around how useful a tool revision is.
I'm impressed how this comment covered all the bases. I would trust your code. haha
Would this be a good time to post about my AI controlled typewriter project?
A computer controlled typewriter is a thing already. It's called a teletype (tty for short).
This line jumps out to me:
> Everything slows down. It’s like back in the old days when you really did one thing at a time.
Why did we turn computers into frenetic, distracted multitasking machines? What would it take to reverse course?
> Why did we turn computers into frenetic, distracted multitasking machines?
When we had to add another core to keep up with Moore’s Law?
I like this. Related, this semester I've been using handwritten quizzes in class. A simple change that's been one of the best things as it changed students' expectations of class prep. Kind of do the readings and sort of prep and you can coast in class. But if you need to write out quiz answers you're forced to know the material better as well as maintain the ability to express yourself.
I also use low-point bonus questions to test general knowledge (huge variation on subjects I thought everyone knew).
I’ve been typing for years since the 80s. However, even in the 90s I found any extended period of handwriting to be painful and laborious. I don’t think I could handle an instructor that insisted on handwritten long form but I’d happily accept a compromise in the form of a typewriter.
Sorry to say you can't take my class.
Does the institution in which your class occurs not have special access procedures? e.g stroke or permanent hand disability
Is there really much point though? I think AI will keep improving, and there will be more and more incentive to use an AI which costs $20/month, instead of a human writer that costs $30/hour. If someone want's an article written, and if people like the AI article as much as the human one, what stops anyone everyone using AI?
The only answer I can think of is that people must believe AI writing will stay below human level for many years, but if so why?
OP is about a German learning class, so yes, there is a point. AI improving will not learn you a language faster than immersing in it until, I suppose, neural links become mainstream.
i dont think that way, ai will became better, but human-taste writing just feels different. like hand-made furniture vs factory-made furniture. they have different class.
I think the AI writing becoming better means it will appear more human rather than like better AI writing. I think the difference in feeling is similar to early attempts to generate faces with AI, which also seemed wierdly wrong in ways which were hard to describe, but now it's very hard to tell them apart.
I like open note exams (and perhaps open book exams, as you need to know the book well to know which page to look at) - it forces you to condense the material to the salient points and operationalise it to solve what would be more challenging problems than a simple recall exam.
When I see 'cheat sheets' - designed to be hidden on the back of calculators or whatever - then I see true application of human ingenuity and intellect.
My school couldn't afford typewriters in the 1980's and early 1990's.
We wrote assignments by hand using a pencil or pen.
Is that really complicated?
When I got to college and everything had to be typed I still wrote everything by hand on paper and edited with an eraser and a red pen to reorganize some sentences or paragraphs. Then I would go to the computer lab and type it in and print it out.
I sometimes wonder if the next step for expensive, elite universities will be to make exam-by-interview a thing, even for undergraduate work. Get in a room with no electronics allowed, just you and the instructor/TA/whoever, and they interview you. Pretty quickly they are going to know if you learned the material or not.
I think if your university doesn't do in person exams with pen and paper then the degrees it hands out are not much evidence of anything.
If you're not interested in learning the course content, then what are you doing there? Pretty expensive waste of time.
I very fondly recall many of the course I did at university. The exams were a helpful motivating factor even for the interesting courses.
This will only work until somebody figures out how to connect an AI to the typewriter which will have some sort of MIC, and the person will start dictating into it with AI-assisted revisions. Once the dictation is over, the AI-enabled typewriter will be instructed to type the work out.
Testing and instruction should be modified to account for AI. If a student uses an Agentic AI for work, learning, research, then when test time comes, the student should be required to stand in the front of the class and teach the class what they have learned, i.e. "Teach Back" all they learned to the entire class student body and teacher. The entire class, instructor included, will also be required to participate in a Q&A session to make sure that student's learning is not just made up of memorization, e.g. restate the information learned but using different words, different scenarios, etc.
FWIW my Dad taught me how to type at 4yo on a huge Imperial typewriter. My spelling took an enormous leap in capability in a few weeks. Primary school teachers were amazed at the words I could spell correctly. (Didn't help my handwriting though which was still like intoxicated chicken scratch on a good day).
Aside from AI-proofing, IMO there is value in slow typing or writing. We have to think just a little bit more before putting ink to paper. There is also a higher cost to making mistakes.
As a kid, before my family could afford a home computer, I was determined to do something that resembled programming. I borrowed "BASIC Computer Games" (1978) by David Ahl[1] from the library and typed in several programs on a manual Olympia typewriter. More than just reading code and maybe even more than being able to easily execute it, I'm convinced this typewriter exercise forced me to really study the flow and the how of the code.
[1] https://archive.org/details/ahl-1978-basic-computer-games/
In college I had to handwrite all my exams and the pure math courses didn't allow calculators. For the latter you could realize you were doing it wrong if the answer was too complicated to write down. As others said the final was something like 30% or more of your grade.
If you need a typewriter, there's a company in New Jersey that makes them for the prison trade in a lucite housing to prevent prisoners from hiding contraband inside. Not sure if China manufactures new machines. India supposedly still manufactures them.
Swintec
College education is all kinds of backwards if the goal was teaching students. If you aren't speaking day one of German find another teacher. In this case you are cheating yourself out of an education if you use a typewriter.
We have amazing tools through screens, but insist on using them as if they were the old tools and act surprised when they lead to worse outcomes. Each student could easily have a friend in Germany they speak to in real time if people actually cared about breaking down cultural barriers. Schools have never cared about education, their main purpose is control. Control of the narrative and control over your life.
When I was a kid we were told no calculators because when you grow up and are in the real world it's not like you will have a calculator with you at all times. Fast forward to today and we all have a calculator on our phones.
I think AI should be treated the same. Who cares if it assists in a lot of the work that is a good thing. BUT as we all know AI has been incorrect on many things so I think what would be a much better learning practice would be to forget if AI wrote the paper and focus heavily on students backing up their claims with sources. So if your paper says ABC is true and AI writes it up in a perfect paragraph you would still need to confirm the facts as true and find a reputable source that shows it to be true.
If I was a professor, I would make a very clear policy they AI is not to be used on assignments, and would repeat it throughout the semester, but make no effort to actually enforce this and even make it easy to abuse.
Then, for the final exam, drop the bomb: in person, handwritten, no outside references, mostly the same assignments we’ve done before. If you fail, it’s over for you. If you stayed true and studied, it should be easy for you to pass. If you used AI all semester, you did it to yourself. Those who complain will have their past assignments audited and if AI was used they are reported for plagiarism. This will be the most valuable lesson.
Soon on Show HN: I built an open source tool that controls your electric typewriter.
Surely there's a middle ground? Get some old wordstar capable 86 class clones and leave the GUI off. It's typing on a keyboard without the confusion of the internet or clicking on glowing icons.
Or just chromebooks with a web address blocklist.
Better dust off that old AlphaSmart!
You could just as easily have a bunch of old desktop computers with no network interface uplink going out of the room active whatsoever, running some basic GUI and libreoffice, connected to nothing but a dumb copper ethernet switch in the same room and an old HP laserjet with a 10/100 ethernet interface in it. No need to force people to deal with typewriters.
> Most students found their pinkies weren’t strong enough to touch-type, so they typed more slowly, pecking at the keyboard with their index fingers.
Huh. I'm not sure I ever use a pinky while touch-typing, except to hit right-backspace sometimes.
For that matter I don't home using F and J either -- I usually home with alt+tab / cmd+tab and right-ctrl / right-cmd.
A typewriter is extreme. In school I used an AlphaSmart (https://en.wikipedia.org/wiki/AlphaSmart, because my handwriting sucked). A laptop without internet would also work.
It's not hard to deploy laptops with all ports and networking disabled, its a BIOS option on most fleet systems.
After 30 seconds the site shows a fullscreen popup:
> The Sentinel not only cares deeply about bringing our readers accurate and critical news, we insist all of the crucial stories we provide are available for everyone — for free.
Thank you very much for interrupting and ruining my reading experience of your article.
What? And after all that money you paid them? The nerve!
Your worldview sucks. Essentially claiming that in order to receive information, you must also receive garbage useless information to brainwash you.
No, complaining about something offered for free because of a minor pop up is what sucks. It's a bizarre sense of entitlement.
If you don't like the website, simply don't use it. Especially when you're making no contribution to it.
I don't think that complaining about things necessitates believing that you're entitled to them. I agree that complaining about things you received for free is in rather poor taste, but I don't think that it's morally wrong in the way that you seem to think it is. If an article you read for free had a pop-up ad on it, you have not been wronged in any way and do not have grounds to sue them, but you should be permitted to voice your complaint, so long as it's of the form "I don't like this" and not "look what they subjected me to, those monsters".
Just because you expect someone to owe something to you, doesn't mean they think they owe you that thing.
If someone gives away something free, they can and sometimes do wash their hands of it. That doesn't prevent you from expressing your opinion on what you think they should change about the work, but they're not under any obligation to do anything about it.
Someone made a thing available. You can take it as it is, you can make noise about what you don't like, you can make it better, or you can ignore it and move on.
If someone is providing a mix of useful and garbage information, well, take your pick from the above.
AI isn’t going anywhere, so blocking it with something like typewriters doesn’t really prepare students for reality.
Feels better to design assignments where students have to use and think about AI, not avoid it. Good experiment, just not a long-term fix.
Remembering my college typewriter-use-by-quarters (coins) on a timer like being at the laundromat, I kind of love this.
At UT Arlington in the Stone Age we had a typewriter lab so folks without home computers with printers could still produce their papers typed, which was required. I had to get a roll of quarters ($10) to do a single paper. And the erase tape was always so used up it was useless.
It was one of the most sadistic things I remember about my college experience, trying to type on those crappy typewriter on a timer. With no errors. And I literally wrote it by hand before trying to transcribe it.
Good luck, we’re all counting on you.
To me this nostalgia is pointless. AI is here and it's good enough only going to get better. The classroom should be about using AI better not ignoring it.
But that would require the teacher to be good at AI too. I think that's the problem here.
> The classroom should be about using AI better not ignoring it.
No, it shouldn’t. I’m not bearish on AI but it shouldn’t replace any part of a classroom where the objective is to learn and communicate in a new language (German). The typewriter argument is memorable and interesting - the article points out the lack of editing forces kids to slow down and think about their writing, as well as iterate through multiple drafts. It’s not a nostalgia thing, they’re not old enough to have ever used one before.
I could see an argument for adding on a new class for GenAI, agents, context engineering or what have you, but considering how behind current US curriculums already are and how quickly the AI field moves, I can only see this ending in wasted time and money: even an up to date class will be stale by the time it’s over. Kids will end up learning this anyway outside of the classroom, no use lecturing them on something they’ll already know.
I completely disagree with you.
You can do all the same learning with AI tools, with the added benefit of developing extra skills on top.
This notion that AI automatically reduces learning seems more born out of fear than reality. There are also different kinds of learning.
People who really want to can still go back to typewriters. Forcing it into a mandatory curriculum though is a step backwards. Just like it would have been pre AI.
Just because an AI can craft a decent Japanese text doesn’t mean I can. Just because AI can write x86 assembly also doesn’t mean I can.
You don’t give first graders a calculator because they will always have one in their pocket- they end up just inputting numbers in a magic box and not learning how to do this manually which will destroy their future mathematical education. It’s about the same with AI.
Here is a good video I watched just yesterday on how to integrate AI into a writing and thinking process.
AI is not a gun that you can't put into the hands of a child. It's a paint brush.
https://www.youtube.com/watch?v=jbHB-rzKBAs
Seems like you're equating 'education' with 'employability training'.
A hand-written essay in class would seem to be a workable mechanism for a student to demonstrate an ability to reason on their own about a subject.
One of my best college professors would review such essays in-person, one-on-one twice each semester.
I spoke with a bunch of profs about how they were assessing students in the age of AI:
https://austinhenley.com/blog/aihomework.html
If students cheat they hurt only themselves. Make sure they understand the consequences for cheating (missing out on learning) and that's about all you can do.
Depends on your measuring stick. Cheating themselves out of an education? Yep. Cheating themselves into a credential -> job - the status / remuneration of which is almost entirely divorced from the quality of the education, being aligned rather with the name of the organization on the diploma.
Former (second-generation) college professor, here. I find it almost impossible to be cynical enough about the US education industry.
The fact that it's an industry is alone enough to cry.
> If students cheat they hurt only themselves
This statement is more defensible after removing “only”. If it “only” hurt the cheaters, there would be no need to police cheating at all.
Well from a certain perspective they are also hurting the schools reputation, the programs reputation, and ultimately their fellow students.
When i attended university (almost a decade ago i guess, time flies) we didn't have a single exam on the computer. All exams were on paper or oral, most were without notes too. Computer science does not require computers.
The thing is, when colleges don't test students' ability properly before issuing a credential, employers start testing job applicants' ability after they've received it.
And they'll do it with all the 'unnecessarily high stakes' and 'risk of unconscious bias' and 'not truly representative' problems that written exams have; and a bunch of extra problems too.
This is usually true, but it is also true that some classes are graded "on a curve" and so grade inflation could hurt people who are honestly doing work. Also, cheaters tend to suck all the air out of a room. For example, my I.T. instructor designed a really nice oral quiz slide-show for the entire classroom. I found it a few hours before the class, I watched it in its entirety, and then when he tried to run it live, I spoilered all the answers before any other student could answer. I wasn't strictly cheating, but I wasn't being fair to my classmates' learning process, either.
They hurt other students who worked hard for the degree. They hurt the reputation of the school and the utility of the degree as a credential.
This is untrue. Students who graduate without actually absorbing knowledge as laid out in the curriculum devalue the degree when they show up in the workforce lacking that knowledge. This is part of why new grads are undesirable job candidates, there’s a chance you are paying a higher wage for someone who may not have learned anything.
Anti-Dvorak instructor... I would do a 360 and walk right out of there. Ain't no way I'm re-learning QWERTY.
Things like this are well-intentioned but idk why there aren't more teachers creating optional "side quests" like these for students that want them instead of forcing them to do things like these
optional "side quests" would allow teachers to create some standard accepted "main quest" curriculum and then just create a bunch of (even possibly "fun") "side quests" students can work on in their spare time for extra skill development
Let me be blunt: the only reason I see not to implement solutions like this appears to be laziness from instructors.
May as well just used pencil and paper. They would have to learn to improve their penmanship too.
this just reinforced my recent belief that the best way to handle the AI boom is to pivot into robotics
I hope they don't force people to walk to school, hunt for food and write in runes.
Typewriters go a bit far, IMO.
Give them a 5150 PC with WordStar and let them go to town.
If AI can do the work, maybe the test should be more focused on what AI can’t do? This is like anyone still doing a traditional coding interview with leetcode problems just because they haven’t yet done the work to figure out what to test for in a world where Claude Code exists.
The goal of the educational process isn't the test paper, it's the learning.
Gyms aren't redundant because tractors exist.
Gyms are a great example actually because tractors exist to do the economically useful work. You now optionally go to the gym to benefit from fake labor that used to be the side effect of useful work. The fake labor is now what colleges are trying to sell, and it's going to kill them.
Gyms predate tractors.
3,000 years ago, physical labor was a component of most jobs. Today gyms are for people who can afford to attend them and don't have a day job that naturally exercises them through labor. People exercising purely for health benefits, and not because the strength benefits them in their job and in other facets of their life, is new.
[dead]
Huh? The gym analogy doesn’t even make sense. People didn’t go to gyms when they were farming with oxen. Gyms are popular now precisely because tractors exist and you don’t need manual labor to farm anymore but people still need the physical exercise for their health. Society has adapted to the arrival of new life-changing technology. Our education system needs to adapt to new technology like AI too. You can probably uplevel a lot of courses and cover a lot more interesting topics than before and teach real application of things you learned aided by AI. Just like when I was doing a CS major 20 years ago, they didn’t spend too much time teaching me assembly programming beyond 1 or 2 lectures (they let me use a compiler for programming assignments!).
Gyms predate tractors by a couple of thousand years. You should think harder about the analogy.
There are plenty of things AI can do that students still benefit from learning.
Maybe instead of trying to teach around the abacus, we need to teach the higher level things you can reach with MATLAB.
We're doing these students a major disservice making them live in the old world. It's our fault for being inflexible, but their world is going to be wholly different and we should just embrace that.
This is like saying you shouldn't learn to add because we have calculators.
I wonder if the daisy wheel printer market just ticked up on eBay.
Just have them write it out. “Ain’t nobody got a goddamn typewriter”.
Next on Show HN: I built a voice activated typewriter.
A typewriter tty would be a fun weekend project.
I’m confused about too many things being measured at once. Is Phelps banning AI to ensure her students are fit to pass terminal examination? And doing so to ensure that her class has a good pass rate, proving she is a good teacher and can keep her job? What if her cohort are particularly dumb? Is she incentivized to make it easy to pass her classes to get that A you paid so much for? Or hard or make that A worth something?
My mentor, a PhD in classics, told me it was never about outcomes and only about improvement. I suppose that answers my question. If your AI gets you an A at the start of the course and an A at the end, then, in the sense that you have not succeeded over anything, you have failed.
My impression was she just brings the typewriters into class as a one-day novelty thing per course, not that it becomes the norm for the whole semester. The goal is to give the students a taste of what the old-fashioned way is like, to get them thinking about it.
Why didn't she ask to write by hand?
Because trying to read the handwriting of 100 students would take a lot of time?
Pfft, just grab a teletype and run lpr -P ttyUSB0 ai_generated_report.txt ;-)
you won't always have a calculator in your pocket
Missing the /s? Most people have a calculator in their pocket all the time now.
I remember having to split bar tabs by hand at the end of the night without a calculator. The waiter delivers a hand-written bill, and we'd calculate how much each person owed, splitting drinks individually and food equally, recalculating tax and tip to derive each person's total. We'd pool the cash in the middle of the table (with people taking their own change as they go where possible), then counting all the cash and making sure it still included sufficient tip. All after many pints. I do NOT miss those days.
Next up: allow slide rules on exams.
Were they ever banned?
Probably around the time they were invented. They were mandatory on my ground exam (private pilot).
OOC was this a while ago? Even when I took the ground exam around 10 years ago, everyone had electronic flight computer calculators (CX-2s).
It was awhile ago (init var me == old;) - back in the era of "iPads can't be used for critical flight information, they're too unreliable".
That makes sense. The CX-2 calculators are a bit less like the iPad era and more like the equivalent of calc I/II classes which only let you use specific TI models versus an app on your smartphone.
It reminds me of a family friend who's a bit older and did their scuba certification using dive tables, whereas when I did my PADI, I was able to use a dive computer.
... and the price of daisywheel printers is skyrocketing. https://en.wikipedia.org/wiki/Daisy_wheel_printing
"You won't always have a calculator in your pocket"
"Orders of magnitude".
Uni professor here.
My colleagues that teach hard skills courses (like data structures and algorithms) either love AI and incorporate it into their teaching at every moment possible, or despise it in the same way graphing calculators were by high school math teachers when they were introduced nearly 30 years ago.
I teach soft skills classes to engineering students, and I'm unconcerned with students using AI. I write my problems in a way such that, if the student truly understands the assignment, prompting the AI to solve the problem and iterating on it takes a similar amount of time to doing the work themselves. AI is not very good at writing introspectively about the student. In other words, AI isn't going to be helpful when the homework question is "A fellow student comes to you asking for suggestions on how to maximize their chances at landing an internship. What advice do you give them that's immediately actionable?"
Try it, plug that into ChatGPT or your favorite LLM. It parrots the same generic tips everyone tells you, with very little on "how" do perform the action in an effective way. Read it, copy it into your advice document, get a poor grade. Try telling other students to take this advice. Note how they don't because the advice isn't actually actionable enough for them to take action.
LLMs are also not very good at the follow-up question "In a previous assignment you gave specific and actionable advice to a peer on the job search. Which of these suggestions were so good you are now doing them?" A number of students write a "Mental Gymnastics" essay, claiming they are following all their suggestions (because they think that's what the professor wants to hear) while the evidence they provide demonstrates they are not. A student asking an LLM to write the essay for them consistently produces a digital 'pat on the back'; a mental gymnastics essay that ultimately makes the student realize how unwilling they are to solve the #1 problem in their college career.
I've done away with exams wherever possible. I stick to project-heavy courses. What I've found to be far more concerning than AI use is the increasing loss of social skills and ability to cooperate within the younger generations. The number of students who would prefer to fail a class instead of talk to literally any human being is astounding.
The number of students who refuse to build soft skills, and believe that tech is truly a meritocracy where the only thing that matters is 'lines of code', there's no politics, and they won't work call or crunch or give code reviews, is also astounding.
... meanwhile, all these students graduate, can't find jobs and become plumbers or bricklayers.
i mean, you can just have AI still do the work, you’re just doing data entry with a type writer.
Doing the data entry might mean you actually learn something. Fairly sure that writing by hand has this effect to some degree.
[flagged]
[flagged]
[dead]
[dead]
[dead]
Might be an unpopular opinion in this thread, but college was made worthless for most degrees as soon as the internet got popular and silly performative shit like this is the death knell. College is about learning how to work in an industry. I'd predict an uptick in trade schools and other hands-on work like medicine, and a continuing downturn in so-called formal education for anything white-collar, programming included. Students are customers. Businesses are going to use AI going forward. No reason to waste time on this.
> College is about learning how to work in an industry.
Oh
Education is a nice side effect sometimes but yeah, I don't know how you could reach any other conclusion. If you're motivated to learn for learning's sake, college is an annoying slog that you know you don't need post-millenium. I literally left college early and started making money instead of spending it, because I got tired of demonstrating to my professors that I already knew everything they were teaching and that it'd be a waste of time for me to come to class.
Or maybe you chose to waste your time because you treated college as a way to get a piece paper instead of as the only time in your life when you are surrounded by experts who will spend an hour a week answering any questions you can think of.
No time wasted at all, that option is also trivially available outside of college, it's called "email". There's a whole industry in tricking new adults into believing that college is not about getting a piece of paper, it's gross, and it's avoidable. I paid off a year of unnecessary college debt in 1/4 of a year of doing real work I learned how to do in my free time. It's a trap and articles like this where colleges are working as hard they can to make education less useful prove it.
>No wasted time
You just said that it was a waste of time. So was it or not?
> that option is also trivially available outside of college, it's called “email”.
How many experts have you cold emailed over the years and how much of their time have you taken?
It would have been more wasted time had I continued after a single year. I went to my first year of college on the advice of my well-meaning parents who are old and like most old people thought it was still important, and yet they agreed with my decision to leave after the first year on an offer for a real six-figure job because there was nothing to learn that I hadn't or couldn't have learned on my own. At least one of my own professors also openly wondered why I was there at all.
To your second question - less than a hundred, but tens. Most people who are worth listening to publish their work and their thoughts. Email is free. Experts love to answer questions about their work, professors hate doing extra work for no extra pay. The incentives here are not confusing. How much time have I taken? Confusing question. These are real people with real passion, and they answer questions with that in mind. Professors are obligated to puke up an answer. I've gotten responses in most cases, in some I haven't. When I don't get answers it's because the targets are smart and busy. If I wanted more engagement with my random questions I'd offer money, and if I had offered money every time I'd still be below par on the money I wasted on college. If I wanted to justify it - I'd say I learned enough to validate that paying real money for another 3-6 years would have been less valuable than burning it for heat.
> At least one of my own professors also openly wondered why I was there at all.
I think you completely misunderstood this interaction.
There are 2 possible explanations.
1. You are so smart/knowledgeable that the professor thinks you are beyond college.
2. You were acting like such an arrogant know-it-all that the professor was being sarcastic.
I’ve seen #1, but I’ve seen #2 many times.
You sound like you have a huge chip on your shoulder about not having a degree. I had the same issue at one point before I went back and finished (after working as a professional developer for a while), so I recognize it.
When I did go back, I asked questions in class, I went to office hours to ask questions, and I did research projects with professors. Some back of the envelope math says it would have costs me about twice what I got out owing if I’d paid for an equal amount of time with whatever experts I could find.
My strong suspicion based on the few posts I’ve read is that your attitude is the reason you had such poor interactions with instructors.
I had excellent interactions with my instructors. I interacted with them like human beings and they understood that their limited time would be better spent with students who didn't have the same energy I did. Several professors, when asked, put me through an impromptu whiteboard quiz and said yeah, do your own thing. It's great that you participated in the process in your own way. In my case I asked if I could show up for the final tests and nothing else, because the intermediate work would have been useless, received permission, and passed.
Chip on my shoulder - no, and it's a silly label to begin with. Understanding that it's for other people who value the paper more than intrinsic understanding, yeah.
EDIT: I will concede in some way that I'm proud of not having a degree, and it does influence my thoughts on this topic. I've met some real idiots that do, and I don't consider it a serious differentiator.
Also looking up the thread - at my early jobs, I was surrounded by many people who were interested in educating me on any topic I could think of, because similarly we were all being paid for our time. The difference between that and school was the assumption that we were both motivated and capable.
1. There’s already a process for testing out of general education classes. You show up, pay $50, and pass a test. You could have saved most of your first year of tuition.
2. These classes that you blew through weren’t upper level classes. They couldn’t have been because you wouldn’t have had the prereqs to take them. If you already had some knowledge of the field and didn’t need lower level classes, you could have talked to the department about testing out of some of them.
I know you didn’t walk into an upper level class on Automata theory and come up with the proofs on the spot.
No professor would in good faith tell you to go do your own thing based on what you’re describing.
If they thought you were very smart and sincere about learning, they’d encourage you to do independent study with them, do research, work with the department to move into higher level classes, or take cross listed graduate classes.
If they thought you were kinda smart, but a huge asshole, they’d tell you to go do your own thing because they didn’t want to deal with your crap.
This is all coming from experience as someone who came into school not needing the intro classes, and someone who used to be that arrogant.
>only time in your life when you are surrounded by experts who will spend an hour a week answering any questions you can think of.
Why is that a resource of "once-in-a-lifetime" scarcity in the first place?
Because
It’s really expensive, time consuming, and difficult to gather exports together like that.
The closest anyone comes to research university is national laboratory, or something like Bell Labs. But you’re unlikely to have access to those.
And you’re unlikely to ever have another time in your life when you can take 4 years to devote almost solely to learning.
Yes exactly, why are those things the case?
It’s expensive because in a capitalist society experts want to be well paid.
As to why you’re unlikely to have another time in your life when you can do this, that’s also because of money. Life tends to get more expensive the older you get.
That doesn't work in tech based professions. In college I took music technology. It was 2 years of my own learning and explaining how everything worked to my tutor.
If you’re talking about a purely vocational program, you’re not likely to be dealing with world class experts.
The college instructor might as well ban calculators and use abacuses then.
We couldn’t use graphing calculators on calculus exams. There were professors who banned calculators entirely.
I'm confused at these comments because it makes me think the commenter didn't go to college. All my classes were basically open everything except for a literal laptop. Calculators were not useful.
at my university in math exams we were only allowed to use 1 specific model of calculator, and most of the exams were answered symbolically anyways, so the calculator usually was not helpful anyways.