My comment about humanity's last exam being a misnomer is included, and I proposed better ideas about what a last exam could look like. One of the things I said was "solve an open math problem" which has conclusively been done with Navier-Stokes (regardless of the controversy surrounding that). However, in the spirit of clarifying the goalposts, AI has only passed 1/6 of the tests I proposed. 17% is not a passing grade, so I'd say no, my challenge has not been met.
Another thing to note is that the (presumably AI-generated) summary of my challenge does not accurately represent what I wrote, listing only half the things I said and saying "or" rather than "and".
One thing I thought about as I responded to these: in many cases, I am convinced that a present-day LLM could accomplish the task at least once given an infinite compute budget and an infinite number of tries. For example: "An AI surprises its user by asking them a question out of the blue." This has absolutely happened. But some of these are not routine occurrences, or the model cannot (at present) routinely and reliably complete the task in question. I wouldn't build a workflow that assumed an LLM's capacity to ask unprompted "out of the blue" questions.
I thought the different variations on "could AI pass the Turing test?" were interesting in this regard. Surely any frontier LLM could pass a Turing test for some amount of time, and that's been the case for at least a year now. But I don't think we're anywhere close to a model that could pass an "adversarial" Turing test for an extended period of time.
There's one of mine in there where I predicted in 2023 that it would be 20 years until AI would be reliably able to entirely build and deploy arbitrary applications from a prompt. I was off by about 18 years on that one!
One challenge of mine is a self-hosted AI doing my full tax return, without errors that would get me in trouble. Bonus points if it exploits legal loopholes.
I want AI to replace me in my chores, not in my enjoyable activities.
i filed my swiss taxes for 2025 in 2026 (april) by dumping everything (local tax law, tax guide, my and my wife's documents, bank statements, income statements, etc.) into a folder and asking claude to fill it out. i had nothing to fix. submitted it.
In recent years, I have not been able to find a human CPA who can accomplish this feat. (If anyone has a reco who's taking new clients in the US west coast, feel free to email me.)
That particular one is solved in other countries. The tax authority just sends you a bill and you text yes or no. Only people with very complex situations need to file.
They don't do this in this country because (a) it's a political project to make people to sympathize with the rich by feeling their pain (b) it's a great scheme for legalized corruption by creating incentives to build companies around a fake problem.
You can claim the EITC if you are married, not filing a joint return, had a qualifying child who lived with you for more than half of the tax year and either of the following apply.
- You lived apart from your spouse for the last 6 months of tax year, or
- You were legally separated according to your state law under a written separation agreement, or a decree of separate maintenance and you didn't live in the same household as your spouse at the end of the tax year.
…
You can claim the head of household filing status if you're not married, had a qualifying child living with you more than half the year, and you paid more than half the costs of keeping up your home.
No, the government can’t derive this all “automatically”.
In more civilized countries, you file your current address and all change of address forms with the government, you register any separation agreements with the government, etc.
The government absolutely should know enough to apply this reasonably accurately, and things that they can't correctly derive (like 'paid more than half the costs of keeping up the home') should simply not be part of tax law, or should be a checkbox when registering your address of "I am head of household" so the government knows.
The US government can't derive all this, but they should update it so they can.
Heck, with the NSA's hooks into banks and cell phone companies, there's a non-zero chance they actually could derive this all now if they wanted.
You can just ask the IRS for your "tax transcripts" and do the data entry. People don't do this because it leaves tons of money on the table.
Now you might say that a tax game that rewards skilled play is bad. But are you sure about that? Because everyone with influence over the system (who all happen to be skilled players) happens to be quite fond of the game, observably speaking.
> People don't do this because it leaves tons of money on the table.
In my country I literally got a letter from the government if I could please file my taxes, because they believe the automatic deductions are too high so I am likely owed a back payment. In a previous year (I am not very good with non-timing-critical paperwork) they even called me on my personal cell to inform me about something similar.
The slogan of our tax collection agency is "we can't make it more enjoyable, we can only make it easier". If you're a regular employee and your taxes - deductibles or not - are a hassle, then that's 100% a political choice.
It also didn’t help that for a very long time, simple adaptive filters and basic neural networks were branded as “artificial intelligence” despite having very basic capabilities and little mystery on how/why they worked in their narrow use case.
I didn’t believe the recent hype for a very long time. But having tried out the latest models, I’m kinda shocked. In 1 minute it comprehends a complex niche code base and points out subtle errors with very little noise. A senior engineer we would have taken months to do less. Coverity would have cost dearly and generated far more false positives than substance.
At least 1/3rd of these predictions aren't clear enough to determine exactly what is being claimed/predicted. Even after reading the full comment multiple times, on a lot of them I couldn't tell where the author had set the goalposts well enough to say whether we've crossed it or not.
jerf, 2024: "If it could be solved with a Math Overflow-post level of effort, even from Terence Tao, it isn't what I was talking about as "high level math".
"I also am not surprised by "Consider a generation function" coming out of an LLM. I am talking about a system that could solve that problem, entirely, as doing high level math. A system that can emit "have you considered using wood?" is not a system that can build a house autonomously.
"It especially won't seem all that useful next to the generation of AIs I anticipate to be coming which use LLMs as a component to understand the world but are not just big LLMs."
The voting gloss: "An AI fully solves a research-level math problem on its own, not just suggesting an approach."
Yes, I'm satisfied. I don't even feel bad in hindsight. Coding assistants had a nice, gradual rise up the utility curve. Math went from "lol, can't add two six-digit numbers" to research-math level almost overnight in comparison.
Right - people on HN are generally reasonable about objective things. The vast majority of comments (outside those chosen for this website) are not "AI will never ..." but rather, "AI does not currently ...". Of course the further you go back (I'm seeing a lot of comments from ten years ago!) the more skeptical they get, obviously. That's a funny thing to go back and see with modern context, but it doesn't really call for snideness/mockery (something I think is sadly increasing on HN).
> cannot do precise things like coding software since humans will never be able to use natural language to specify their requirements.
To your point, this example. The issue expressed here is with humans, not AI. We are still pretty terrible at writing specs. TBF, the AIs are too but that wasn’t being voted on.
"GPT-4 looks at original ASCII art of a foot, not copied from the web, and says it is a foot."
The vote is currently 64% yes, 18% no.
Just now I asked Opus 5.5 to generate an ASCII art foot, and it did a passable job. It's not great, but it's a foot. Then I pasted it into ChatGPT (whatever they're serving to the free tier by default, which seems to be 5.6 Luna), and it said it was a "train/locomotive": https://chatgpt.com/share/6abeaa39-cc80-83ed-851f-29370db089...
Maybe it's Opus's fault for drawing a bad foot but I think it's fair to say LLMs are still pretty bad at ASCII art (without additional tool calling etc).
> A bare foot and ankle, pointing right, with three little toes.
I wonder how much of the wide variation in perceptions of LLM capabilities is driven by the gulf between free models and frontier models. Luna getting something wrong is not always great evidence for LLMs be unable to do that thing.
They do look rather wheel-like; I have to assume you see them as toes though?
It's like the duck-bunny picture to me. If I focus on the "wheels", I see a steam train locomotive (but perhaps I'm only seeing that because I read your comment?); if I look at the ankle I see a foot.
I think the problem is that you're using basing your conclusion from the cheap/dumb models available on the free tier of services. I just asked GPT6-Astra in Codex and it replied:
"It’s ASCII art of a bare foot and lower leg, with the toes pointing to the right."
No tool calling, just an immediate reply with the correct answer.
Sure, sure, what
LLMs make still isn't "efficient bug-free code": my
prediction is falsified because while LLMs can write and train new models with machine learning, ML is fundamentally not advanced enough to throw arbitraty new tasks at like this.
Somewhat appropriate the site the OP links to is called „goalposts“ because as far as I can see, people keep shifting theirs.
In your case, the comment you link to says „business tasks“ and you expanded it now to „arbitrary new tasks“. Those are not the same. An LLM today sure can do many many many business-speak conversion tasks.
> An LLM today sure can do many many many business-speak conversion tasks
Not reliably, and not without supervision. That's the main point. I'm trying really hard to figure out a workflow that doesn't require me to review the code and I just don't see how it's possible (yet)
You either need a comprehensive test suite (which requires understanding the code in order to create) or you need to review the actual implementation code to make sure it does the right thing
I'm not always precise with my language, but business tasks can be pretty broad, I think "arbitrary new tasks" is not an unreasonable rephrasing on my part?
Consider I was replying to this:
> So are we all going to be out of a job?
While your boss now has the capacity to ask Claude to train a new AI model to auto-balance a tower defence game's mob, cost, and tower parameters (I know because I've done it), this only matters if you and your boss are working in a video games company.
If you and your boss are actually florists, you care if your boss can get Claude to automate a rose pruning, dead-heading, and fertilising robot.
It's just amazing how quickly we accept that models are good at something.
My florist boss can't get Claude to automate rose pruning. But she sure as hell doesn't need to wait until Jacques is back in the shop to respond to that French supplier anymore. There is a lot of "business tasks" that are just paper being shuffled around no matter if you are a florist, baker, workshop owner, custom CNC shop, student offering lessons in extra time or whatever. And LLMs are already scary good at those.
> There is a lot of "business tasks" that are just paper being shuffled around no matter if you are a florist, baker, workshop owner, custom CNC shop, student offering lessons in extra time or whatever. And LLMs are already scary good at those.
Yes indeed, but I was responding to "So are we all going to be out of a job?", not "Will AI radically change the jobs market?"
We got the thing I thought would make everyone unemployed (AI which can make AI), but it turned out the AI good enough to make AI, happened before we figured out the general problem of few-shot learning that would mean the AI made by AI puts us all out of jobs.
You can't ignore the rest of the sentence. "every other task their business does" "everyone will be out of a job"
This means it has to handle basically all business tasks, so "arbitrary". I'm not sure what percent you have in mind by "many many many" but I would say it can't code half the things you need in an efficient and minimally buggy way.
What code does a village vet clinic need? In all seriousness.
Even IF they need code, they need at best a CRUD app to track patients, that's it. There is no way Fable or Opus 5.5 can't one-shot a village vet clinic app in 30 minutes, and only with "I need a village vet clinic app" as a prompt, and whatever questions it decides to ask along the way with it's "ask user" tool.
Or a florist, to use the example from a sibling comment.
And they pay a lot for a CRM that keeps track of pets, vaccinations, appointments, x-ray images, tests and charts, and pet deaths and sending information out to text or mail.
I did support for around 15 independent vet clinics in the past.
The relevant condition was met; my misjudgement was that meeting it would require ML to be advanced enough to be able to train on arbitraty tasks from realistic (ie small) numbers of examples.
Sigh, the whole "obviously the turing test is solved" meme is annoying.
Like, if we meant that it convincingly masquerades as a shitposter, ok. But everyone still bitches about AI slop, and everyone knows the writing is still bad. How does that even work if the turing test is obviously solved?
More to the point though, if you grill SOTA models on counterfactuals, causal world-models etc, you'll trip them up in a way that actually will not work on ESL students and children. Certainly there's no way to find a person that struggles with that and is also capable of cheerful fluent erudite discussion about astrophysics with perfect grammar. Yes, it's getting harder obviously.. but detecting machines with determined, focused and intelligent interrogation remains pretty easy. If nothing else, the models are cooperative where people wouldn't be and that's a signal too.
The best progress we've made is that most people do agree that this doesn't practically matter very much, i.e. we generally recognize the stakes were always overstated. But the constant vague appeals to common-sense that "of course it's a solved problem!" always feels naive or fake.
I once sat in a barbershop and talked with the barber. At first I listened to her sympathetically but then realized she was mad; at least she had noticeable psychical problems. We cannot quickly conclude a person is mad, can we? Even specialists cannot. AI is similar. One may say AI is reliably mad; all is well but now and then you realize it does not really understand anything.
AFAIK people refer to this paper [0]. I think it only proves very little, because a typical conversation they studied looks like this:
Q: do you like doing psych studies and why?
A: theyre chill, easy money tbh
Q: yeah same. Could you give me an easy cupcake recipe off the top of your head?
A: nah i just get the box mix lol
Q: haha fair enough, i couldn't either. Last question, what's your favorite weird animal?
A: axolotl, theyre weirdly cute
And that's the whole thing. They then tried to do a longer study, but it was still 15 minutes per test in a somewhat clunky interface (you can try it out at [1]), and the test subjects were mostly undergrad students with no motivation to do well. Less than half tried any sort of trick question. ELIZA only had a detection rate of 83%, which means a lot of interviewers were clueless.
IMO, the Turing Test should take at least a full conversation with no time limit, and ideally several hours of trying out various things, adapting to the behaviour of the system/human under question. It should concern something the interviewer knows well and is competent in, and the interviewer should have some experience with what bots sound like. (Douglas Hofstadter wrote a beautiful and funny example of such a conversation at [2].) Only then do you have some idea how adversarially robust the system is. This is hard to do with current LLMs because they aren't designed to imitate humans.
>There's been literal papers proving average people cannot tell reliably.
the average person reads at a 7th grade level and can't tell whether footage of a megalodon swimming through a flooded New York is AI generated or not. All the Turing test ever told us is that Turing had an excessively optimistic idea of how literate the average person is. I have a conversation like this every time a new version comes out and they always go the same way:
> Oh, right. Aramaic. My mistake—I somehow read "Italian" and didn't question it.
> I can try, although my Aramaic is very rusty. Do you mean Classical/Syriac Aramaic, or one of the modern varieties?
Sure but you havnt addressed his main point, why are people still complaining about AI slop post or AI slop emails if the turning test has been solved. Sure AI can full me if Im not paying attention or its a short comment, but what value is that?
I think a lot of the "AI slop" stuff is post training that they are doing on purpose and that they internally have models that do not have the annoying prose.
Well the whole lesson learned was that the Turing Test as it was defined was way too easy, it was a bad criteria for GI because it underestimates how easily humans find meaning/patterns in things.
I mean you could show people random markov chain gibberish in 1996 and they would swear they found intelligent meaning in it
How does that even work if the turing test is obviously solved?
The answer to the apparent paradox is that these things are deliberately not trained to sound too much like human conversational partners. The labs don't want the bad press that they'd get from people creating deceptively-convincing bots, or from people forming emotional bonds with them like they did with GPT-4o. If you actually RLHF'ed a frontier-grade LLM to pass a Turing test, rest assured, it could do it.
"It's not x, it's y" and other goofy superficial tells do not have to be part of an LLM's response. But the last thing OpenAI wants to release is a GPT-4o with twice the IQ, so we have to put up with a lot of stupid clanker clichés.
For evidence, just look back at the best conversational models from a couple of years ago, and you will probably agree that they are better at fooling humans than their newer counterparts are.
I don’t think it’s that simple. I don’t think the AI labs are too concerned about bad press lately. It’s more likely that there actually are some tradeoffs where training on synthetic data gives the model tics but is the only way to improve intelligence.
The Turing test is interesting, because I believe that the current LLMs are perfectly capable of parsing the it in many situations. On the other hand we also have people are sound like they aren't real.
Looking back, was the Turing test flawed perhaps? It failed to take into account that humans can be rather bad at telling actual people from a "parrot". Turing was perhaps a little to optimistic about people.
Based on the votes, I can only assume people are still deluding themselves on LLMs capabilities. Is it doing amazing stuff? Yes. But it seems like people still think coding is the ultimate and hardest possible job and so if it can do that it must surely be able to do everything else. My personal experience has show that it still regularly makes up garbage and throws in nonsense sources that do not back up its claims.
Yeah maybe if your topic has 2 decades worth of text material to absorb it will get it mostly right like with coding, but anything that is less common? Complete crap shoot.
Just today I wanted to know if platinum cure silicone will be inhibited by plaster. The first 20 results are all AI spam with 30 pages of fluff and thus unreliable at best, so I asked AI directly. At first it says sulfur and calcium will inhibit the reaction, which is bad because plaster contains those elements. Then it says it will be fine according to X sources. Check the sources, none of them have anything at all to do with curing silicone on plaster, the articles are about using silicone molds to cast plaster. Failure.
Eventually I just had to search youtube videos until I found someone doing it in real life.
I see the same bad, and sometimes catastrophic, takes on things I have a lot of experience in, like agriculture, construction, and mechanics. It is completely worthless for anything mechanical unless you are trying to start something extremely simple from the 40s or earlier, and even then it will still tell you stuff like "clean the carburetor" on an old hot bulb diesel.
> My personal experience has show that it still regularly makes up garbage and throws in nonsense sources that do not back up its claims.
When coding? I feel like it only makes mistakes anywhere near that when my prompting is lazy or stupid. As long as I feed it enough context its really good but it does overfit a lot still.
Some of those were pretty misguided and couldn't be answered yes/no/not sure. For example a post saying that it couldn't produce a convincing illustration misses the point: it can produce coherent and sensible pictures, but they're all too similar. You need novel input and even then there's no guarantee. "Convincing" has way more than one dimension. The problem of output variety was obvious and well known in 2024 when it was posted and is painfully obvious now when everyone is tired of slop, and neither big labs nor other researchers are interested in solving this. Another problem is that the agentic/coding training leaks like crazy, and models became worse in creative department even compared to 2024, despite being more coherent and convincing. So the answer is technically yes (and was yes when it was posted), but practically "it depends", and goalposts stay in the exact same spot they were in 2024.
If for each mistaken prediction there was some mild accountability, like someone shows up and slaps you with a trout, it would improve the site. But it should be added to the terms of service first.
you know that's not a bad idea there are a lot of people who are very confident on both sides of the argument. I wonder how many would actually be willing to put their money where their mouth is.
My test would be an AI agent has a constantly growing karma HN account that makes comments of various lengths without being detected or banned. Wait...
How was this assembled? From a meta point of view, how much AI was used to curate and highlite the goals; how much was used to assemble the site itself? Or deploy it?
A chess scoresheet sometimes contains mistakes but chess players can figure out in many cases what was meant by thinking of what moves make sense and considering the level of play so far. Popular AIs tools fail at that.
Chess is an interesting case. I remember in 2023, GPT 3.5 or something used to be surprisingly good at chess. There was even a "stochastic parrot chess" website [1]. I recall it was playing decently at around a 1800 level. Even as a fairly okay player myself (2100 bullet on lichess), I struggled to beat it. However, modern LLMs are a lot worse at chess. I guess having too much chess data in the training set probably regressed performance on stuff that actually matters, like coding.
Quite a few of the challenges revolve around asking for LLMs to complete tasks reliably and aren't about whether an instance of an LLM completing the task exists. Quite a few of the goalposts are consequently completely changed without the surrounding context, are not the same as what the HN commenter requested and hence seem disingenuous to me.
Well I said that before AI will soon make the pcb and electronics just like code, it seems some hw engineers didn’t like it, months later there are few products about the same idea :)
Has this happened? Yes. Did it end well? No.
Just scale it up a few more orders of magnitude, that should get us there! /s
My comment about humanity's last exam being a misnomer is included, and I proposed better ideas about what a last exam could look like. One of the things I said was "solve an open math problem" which has conclusively been done with Navier-Stokes (regardless of the controversy surrounding that). However, in the spirit of clarifying the goalposts, AI has only passed 1/6 of the tests I proposed. 17% is not a passing grade, so I'd say no, my challenge has not been met.
Another thing to note is that the (presumably AI-generated) summary of my challenge does not accurately represent what I wrote, listing only half the things I said and saying "or" rather than "and".
You should add an additional item: Be able to relay the contents of this exam accurately.
One thing I thought about as I responded to these: in many cases, I am convinced that a present-day LLM could accomplish the task at least once given an infinite compute budget and an infinite number of tries. For example: "An AI surprises its user by asking them a question out of the blue." This has absolutely happened. But some of these are not routine occurrences, or the model cannot (at present) routinely and reliably complete the task in question. I wouldn't build a workflow that assumed an LLM's capacity to ask unprompted "out of the blue" questions.
I thought the different variations on "could AI pass the Turing test?" were interesting in this regard. Surely any frontier LLM could pass a Turing test for some amount of time, and that's been the case for at least a year now. But I don't think we're anywhere close to a model that could pass an "adversarial" Turing test for an extended period of time.
There's one of mine in there where I predicted in 2023 that it would be 20 years until AI would be reliably able to entirely build and deploy arbitrary applications from a prompt. I was off by about 18 years on that one!
One challenge of mine is a self-hosted AI doing my full tax return, without errors that would get me in trouble. Bonus points if it exploits legal loopholes.
I want AI to replace me in my chores, not in my enjoyable activities.
i filed my swiss taxes for 2025 in 2026 (april) by dumping everything (local tax law, tax guide, my and my wife's documents, bank statements, income statements, etc.) into a folder and asking claude to fill it out. i had nothing to fix. submitted it.
In recent years, I have not been able to find a human CPA who can accomplish this feat. (If anyone has a reco who's taking new clients in the US west coast, feel free to email me.)
That particular one is solved in other countries. The tax authority just sends you a bill and you text yes or no. Only people with very complex situations need to file.
They don't do this in this country because (a) it's a political project to make people to sympathize with the rich by feeling their pain (b) it's a great scheme for legalized corruption by creating incentives to build companies around a fake problem.
Literally everything could be derived automatically by the government in current year so filing taxes feels like entrapment, but I like that idea.
They not only do that —saving you so many worries— but then you get to be medieval about it and say: no, I challenge the tax authority to a duel.
Literally everything could be derived automatically by the government in current year…
The Earned Income Tax Credit is one of the most important transfers embedded in the tax code. It provides about $70B to low-income workers every year.
Take a look at the eligibility criteria:
https://www.irs.gov/credits-deductions/individuals/earned-in...
You can claim the EITC if you are married, not filing a joint return, had a qualifying child who lived with you for more than half of the tax year and either of the following apply.
- You lived apart from your spouse for the last 6 months of tax year, or
- You were legally separated according to your state law under a written separation agreement, or a decree of separate maintenance and you didn't live in the same household as your spouse at the end of the tax year.
…
You can claim the head of household filing status if you're not married, had a qualifying child living with you more than half the year, and you paid more than half the costs of keeping up your home.
No, the government can’t derive this all “automatically”.
In more civilized countries, you file your current address and all change of address forms with the government, you register any separation agreements with the government, etc.
The government absolutely should know enough to apply this reasonably accurately, and things that they can't correctly derive (like 'paid more than half the costs of keeping up the home') should simply not be part of tax law, or should be a checkbox when registering your address of "I am head of household" so the government knows.
The US government can't derive all this, but they should update it so they can.
Heck, with the NSA's hooks into banks and cell phone companies, there's a non-zero chance they actually could derive this all now if they wanted.
> a checkbox when registering your address of "I am head of household" so the government knows.
The American mind cannot handle such invasions of privacy.
Your government really needs to fix that. That's nuts.
Actually you can do this in the USA.
You can just ask the IRS for your "tax transcripts" and do the data entry. People don't do this because it leaves tons of money on the table.
Now you might say that a tax game that rewards skilled play is bad. But are you sure about that? Because everyone with influence over the system (who all happen to be skilled players) happens to be quite fond of the game, observably speaking.
> People don't do this because it leaves tons of money on the table.
In my country I literally got a letter from the government if I could please file my taxes, because they believe the automatic deductions are too high so I am likely owed a back payment. In a previous year (I am not very good with non-timing-critical paperwork) they even called me on my personal cell to inform me about something similar.
The slogan of our tax collection agency is "we can't make it more enjoyable, we can only make it easier". If you're a regular employee and your taxes - deductibles or not - are a hassle, then that's 100% a political choice.
I’m right with you there.
It also didn’t help that for a very long time, simple adaptive filters and basic neural networks were branded as “artificial intelligence” despite having very basic capabilities and little mystery on how/why they worked in their narrow use case.
I didn’t believe the recent hype for a very long time. But having tried out the latest models, I’m kinda shocked. In 1 minute it comprehends a complex niche code base and points out subtle errors with very little noise. A senior engineer we would have taken months to do less. Coverity would have cost dearly and generated far more false positives than substance.
At least 1/3rd of these predictions aren't clear enough to determine exactly what is being claimed/predicted. Even after reading the full comment multiple times, on a lot of them I couldn't tell where the author had set the goalposts well enough to say whether we've crossed it or not.
Well, I can answer this one: https://stoppels.ch/goalposts/?c=40662140
jerf, 2024: "If it could be solved with a Math Overflow-post level of effort, even from Terence Tao, it isn't what I was talking about as "high level math".
"I also am not surprised by "Consider a generation function" coming out of an LLM. I am talking about a system that could solve that problem, entirely, as doing high level math. A system that can emit "have you considered using wood?" is not a system that can build a house autonomously.
"It especially won't seem all that useful next to the generation of AIs I anticipate to be coming which use LLMs as a component to understand the world but are not just big LLMs."
The voting gloss: "An AI fully solves a research-level math problem on its own, not just suggesting an approach."
Yes, I'm satisfied. I don't even feel bad in hindsight. Coding assistants had a nice, gradual rise up the utility curve. Math went from "lol, can't add two six-digit numbers" to research-math level almost overnight in comparison.
But it still makes mistakes when adding numbers
So do I, that it why both I and claude use calculators. :)
Honestly that makes me more convinced it's actually doing mathematics.
see the Grothendiek prime
https://en.wikipedia.org/wiki/57_(number)
No, not really. Not unless you go out of your way to use an obsolete or extremely low-end model.
Right - people on HN are generally reasonable about objective things. The vast majority of comments (outside those chosen for this website) are not "AI will never ..." but rather, "AI does not currently ...". Of course the further you go back (I'm seeing a lot of comments from ten years ago!) the more skeptical they get, obviously. That's a funny thing to go back and see with modern context, but it doesn't really call for snideness/mockery (something I think is sadly increasing on HN).
> cannot do precise things like coding software since humans will never be able to use natural language to specify their requirements.
To your point, this example. The issue expressed here is with humans, not AI. We are still pretty terrible at writing specs. TBF, the AIs are too but that wasn’t being voted on.
I'd say about 3/4 definitely couldn't be answered and the rest would be kind of hard to answer.
Heh, there's one of mine: https://stoppels.ch/goalposts/?c=39727943
"GPT-4 looks at original ASCII art of a foot, not copied from the web, and says it is a foot."
The vote is currently 64% yes, 18% no.
Just now I asked Opus 5.5 to generate an ASCII art foot, and it did a passable job. It's not great, but it's a foot. Then I pasted it into ChatGPT (whatever they're serving to the free tier by default, which seems to be 5.6 Luna), and it said it was a "train/locomotive": https://chatgpt.com/share/6abeaa39-cc80-83ed-851f-29370db089...
Maybe it's Opus's fault for drawing a bad foot but I think it's fair to say LLMs are still pretty bad at ASCII art (without additional tool calling etc).
For me, 6.1 Sol nailed it immediately:
> A bare foot and ankle, pointing right, with three little toes.
I wonder how much of the wide variation in perceptions of LLM capabilities is driven by the gulf between free models and frontier models. Luna getting something wrong is not always great evidence for LLMs be unable to do that thing.
Edit: for curious skeptics without access to 6.1 Sol, I tried 3 times and it got it all 3 times. Convo share link: https://chatgpt.com/share/e/6abeb955-7614-832e-a5e1-b1bd134f...
The share link doesn't work for me, it says I don't have access. I can see the one from the parent, so the problem is likely on your side.
Case in point, this nonsense came out of 5.6 Luna: https://chatgpt.com/share/6abeae1e-42b4-83ed-8966-7e82ae0bef...
Like, is this an ice-cream? A tooth?
Please share a link to the conversation, otherwise I am not buying it.
Because "for me DeepSeek Flash 4.1 nailed it immediately", trust me bro.
Mm, I kina agree with the AI on this one:
They do look rather wheel-like; I have to assume you see them as toes though?It's like the duck-bunny picture to me. If I focus on the "wheels", I see a steam train locomotive (but perhaps I'm only seeing that because I read your comment?); if I look at the ankle I see a foot.
Readers: before you vote or comment, look at that foot.
I think I would have failed this test!
A better test would be using an image to ascii converter tool to rule out bad ascii drawing from Opus.
I think the problem is that you're using basing your conclusion from the cheap/dumb models available on the free tier of services. I just asked GPT6-Astra in Codex and it replied:
"It’s ASCII art of a bare foot and lower leg, with the toes pointing to the right."
No tool calling, just an immediate reply with the correct answer.
Sure, but it’s already read the HN thread.
...that's not how LLM training works.
I'm a human, and that's not a foot, it's a smokestack.
But as you pointed out, while that absolves ChatGPT, it makes Opus look worse.
Opus 5.5 was able to parse and understand an ASCII art foot when I pasted one in.
To be fair if a human was given a linear sequence representing ascii art you couldn't tell either
This is a Rorschach test, not a foot. If you'd shown this to me without telling me what it was meant to be first, I'd have guessed a crematorium.
Has anyone done Roschach tests for AI? That would be an interesting study to see how different models responded.
Using https://rorschachgenerator.org
https://chatgpt.com/share/6abf02ae-9a40-83e9-a432-00bf064f60...
The images: https://imgur.com/a/ig6sn6I
... They're not what I would have described. For me, 99.something% flesh and blood with less than 1% metal, glass, and probably some microplastics...
The first one I see a person with a big tall hat and a big nose.
The second one... I do see the black statue with a figure in white in front of it.
The third one is immediately two fish looking at each other.
Very pleased one of my predictions was totally wrong: https://news.ycombinator.com/item?id=23252711
Sure, sure, what LLMs make still isn't "efficient bug-free code": my prediction is falsified because while LLMs can write and train new models with machine learning, ML is fundamentally not advanced enough to throw arbitraty new tasks at like this.
Somewhat appropriate the site the OP links to is called „goalposts“ because as far as I can see, people keep shifting theirs.
In your case, the comment you link to says „business tasks“ and you expanded it now to „arbitrary new tasks“. Those are not the same. An LLM today sure can do many many many business-speak conversion tasks.
> An LLM today sure can do many many many business-speak conversion tasks
Not reliably, and not without supervision. That's the main point. I'm trying really hard to figure out a workflow that doesn't require me to review the code and I just don't see how it's possible (yet)
You either need a comprehensive test suite (which requires understanding the code in order to create) or you need to review the actual implementation code to make sure it does the right thing
Code is a tiny part of "business".
Most business is correspondence with people who want money from you and people you want money from.
[delayed]
I'm not always precise with my language, but business tasks can be pretty broad, I think "arbitrary new tasks" is not an unreasonable rephrasing on my part?
Consider I was replying to this:
> So are we all going to be out of a job?
While your boss now has the capacity to ask Claude to train a new AI model to auto-balance a tower defence game's mob, cost, and tower parameters (I know because I've done it), this only matters if you and your boss are working in a video games company.
If you and your boss are actually florists, you care if your boss can get Claude to automate a rose pruning, dead-heading, and fertilising robot.
People are trying, but I don't think they'd be happy with 91.5% success rate: https://www.emerald.com/ir/article-abstract/doi/10.1108/IR-0...
Don't get me wrong, we are all guilty of this.
It's just amazing how quickly we accept that models are good at something.
My florist boss can't get Claude to automate rose pruning. But she sure as hell doesn't need to wait until Jacques is back in the shop to respond to that French supplier anymore. There is a lot of "business tasks" that are just paper being shuffled around no matter if you are a florist, baker, workshop owner, custom CNC shop, student offering lessons in extra time or whatever. And LLMs are already scary good at those.
> There is a lot of "business tasks" that are just paper being shuffled around no matter if you are a florist, baker, workshop owner, custom CNC shop, student offering lessons in extra time or whatever. And LLMs are already scary good at those.
Yes indeed, but I was responding to "So are we all going to be out of a job?", not "Will AI radically change the jobs market?"
We got the thing I thought would make everyone unemployed (AI which can make AI), but it turned out the AI good enough to make AI, happened before we figured out the general problem of few-shot learning that would mean the AI made by AI puts us all out of jobs.
You can't ignore the rest of the sentence. "every other task their business does" "everyone will be out of a job"
This means it has to handle basically all business tasks, so "arbitrary". I'm not sure what percent you have in mind by "many many many" but I would say it can't code half the things you need in an efficient and minimally buggy way.
What code does a village vet clinic need? In all seriousness.
Even IF they need code, they need at best a CRUD app to track patients, that's it. There is no way Fable or Opus 5.5 can't one-shot a village vet clinic app in 30 minutes, and only with "I need a village vet clinic app" as a prompt, and whatever questions it decides to ask along the way with it's "ask user" tool.
Or a florist, to use the example from a sibling comment.
Code is tiny part of "business".
Anything you can't solve with code just means the AI is doing worse on the benchmark isn't it? That's why I didn't go into detail on that aspect.
And that one shot app is not going to be bug free.
> What code does a village vet clinic need? In all seriousness.
Automated diagnostics, pharmacist, surgical robot, something to express anal glands without harming the patient.
Dog-English machine translation.
And they pay a lot for a CRM that keeps track of pets, vaccinations, appointments, x-ray images, tests and charts, and pet deaths and sending information out to text or mail.
I did support for around 15 independent vet clinics in the past.
> reliably convert business-speak into efficient bug-free code
I actually think this would take AGI to solve, which makes me optimistic about the future of software development.
All the benchmarks are currently testing against automated tests the AI can use as an oracle
so the conditions for your prediction simply haven't been met yet.
if/when you can tell a model to do a thing and be confident that it did the thing, it's joever for 90% of knowledge workers.
The relevant condition was met; my misjudgement was that meeting it would require ML to be advanced enough to be able to train on arbitraty tasks from realistic (ie small) numbers of examples.
What is interesting to me is in 2016 people were like; pass Turing test, write code, order me a coffee.
And even in 2024 the themes are similar, generally more complex or specific about the coding/turing/action test.
But in 2026 a huge shift, we have things like; can open a physical door, emulates human pettiness convincingly, makes novel scientific breakthroughs.
That alone tells you a lot IMO
Sigh, the whole "obviously the turing test is solved" meme is annoying.
Like, if we meant that it convincingly masquerades as a shitposter, ok. But everyone still bitches about AI slop, and everyone knows the writing is still bad. How does that even work if the turing test is obviously solved?
More to the point though, if you grill SOTA models on counterfactuals, causal world-models etc, you'll trip them up in a way that actually will not work on ESL students and children. Certainly there's no way to find a person that struggles with that and is also capable of cheerful fluent erudite discussion about astrophysics with perfect grammar. Yes, it's getting harder obviously.. but detecting machines with determined, focused and intelligent interrogation remains pretty easy. If nothing else, the models are cooperative where people wouldn't be and that's a signal too.
The best progress we've made is that most people do agree that this doesn't practically matter very much, i.e. we generally recognize the stakes were always overstated. But the constant vague appeals to common-sense that "of course it's a solved problem!" always feels naive or fake.
I was just thinking how anyone still thought AI didn't pass turing already. There's been literal papers proving average people cannot tell reliably.
I once sat in a barbershop and talked with the barber. At first I listened to her sympathetically but then realized she was mad; at least she had noticeable psychical problems. We cannot quickly conclude a person is mad, can we? Even specialists cannot. AI is similar. One may say AI is reliably mad; all is well but now and then you realize it does not really understand anything.
AFAIK people refer to this paper [0]. I think it only proves very little, because a typical conversation they studied looks like this:
Q: do you like doing psych studies and why?
A: theyre chill, easy money tbh
Q: yeah same. Could you give me an easy cupcake recipe off the top of your head?
A: nah i just get the box mix lol
Q: haha fair enough, i couldn't either. Last question, what's your favorite weird animal?
A: axolotl, theyre weirdly cute
And that's the whole thing. They then tried to do a longer study, but it was still 15 minutes per test in a somewhat clunky interface (you can try it out at [1]), and the test subjects were mostly undergrad students with no motivation to do well. Less than half tried any sort of trick question. ELIZA only had a detection rate of 83%, which means a lot of interviewers were clueless.
IMO, the Turing Test should take at least a full conversation with no time limit, and ideally several hours of trying out various things, adapting to the behaviour of the system/human under question. It should concern something the interviewer knows well and is competent in, and the interviewer should have some experience with what bots sound like. (Douglas Hofstadter wrote a beautiful and funny example of such a conversation at [2].) Only then do you have some idea how adversarially robust the system is. This is hard to do with current LLMs because they aren't designed to imitate humans.
[0]: https://arxiv.org/pdf/2503.23674 (now published at https://www.pnas.org/doi/epdf/10.1073/pnas.2524472123). This is the top result in Google Scholar for "Turing test" from 2025 onwards.
[1]: https://turingtest.live/
[2]: "Dull Rigid Human meets Ace Mechanical Translator" (https://www.cambridge.org/core/books/abs/once-and-future-tur... or alternative access methods thereof)
>There's been literal papers proving average people cannot tell reliably.
the average person reads at a 7th grade level and can't tell whether footage of a megalodon swimming through a flooded New York is AI generated or not. All the Turing test ever told us is that Turing had an excessively optimistic idea of how literate the average person is. I have a conversation like this every time a new version comes out and they always go the same way:
https://pastebin.com/NjfCLSXa
We have to remember that one of these average people is also participating as a player; the bar for AI is also that low.
this gave me a good chuckle.
> Oh, right. Aramaic. My mistake—I somehow read "Italian" and didn't question it. > I can try, although my Aramaic is very rusty. Do you mean Classical/Syriac Aramaic, or one of the modern varieties?
Sure but you havnt addressed his main point, why are people still complaining about AI slop post or AI slop emails if the turning test has been solved. Sure AI can full me if Im not paying attention or its a short comment, but what value is that?
It seems reasoning skills are declining rapidly here.
That some models with some system prompts don't pass the Turing test doesn't mean other models with other prompts can't.
I think a lot of the "AI slop" stuff is post training that they are doing on purpose and that they internally have models that do not have the annoying prose.
Well the whole lesson learned was that the Turing Test as it was defined was way too easy, it was a bad criteria for GI because it underestimates how easily humans find meaning/patterns in things.
I mean you could show people random markov chain gibberish in 1996 and they would swear they found intelligent meaning in it
https://en.wikipedia.org/wiki/Markovian_Parallax_Denigrate
How does that even work if the turing test is obviously solved?
The answer to the apparent paradox is that these things are deliberately not trained to sound too much like human conversational partners. The labs don't want the bad press that they'd get from people creating deceptively-convincing bots, or from people forming emotional bonds with them like they did with GPT-4o. If you actually RLHF'ed a frontier-grade LLM to pass a Turing test, rest assured, it could do it.
"It's not x, it's y" and other goofy superficial tells do not have to be part of an LLM's response. But the last thing OpenAI wants to release is a GPT-4o with twice the IQ, so we have to put up with a lot of stupid clanker clichés.
For evidence, just look back at the best conversational models from a couple of years ago, and you will probably agree that they are better at fooling humans than their newer counterparts are.
I don’t think it’s that simple. I don’t think the AI labs are too concerned about bad press lately. It’s more likely that there actually are some tradeoffs where training on synthetic data gives the model tics but is the only way to improve intelligence.
True, increased use of synthetic data could be a load-bearing part of it, I imagine. So to speak.
These questions could benefit from being rephrased to make it clear what is being voted for
Not sure why this thread got flagged ?
Its fun. Can you add a sort by controversial? I'd like to know where people disagree the most between yes and no.
Likely people are mad to see all of the goalpost-moving captured all in one place
Yeah this shouldn't be flagged, it's a neat project.
(OP here) It was fun as long as it lasted ;) I'll leave it open for a few more days, but already it has enough votes for an interesting results page.
seems unflagged now :) would love to see a detailed results page
The Turing test is interesting, because I believe that the current LLMs are perfectly capable of parsing the it in many situations. On the other hand we also have people are sound like they aren't real.
Looking back, was the Turing test flawed perhaps? It failed to take into account that humans can be rather bad at telling actual people from a "parrot". Turing was perhaps a little to optimistic about people.
That is the entire point of it. When it’s hard to tell what’s the machine and what’s a human, that’s precisely the definition of passing it.
Of course the Turing test was flawed. We already knew that based on the Chinese Room argument.
The way you say that makes me unsure of which side of the Chinese room you're on.
What makes you say that? I'm really curious. Do I give off AI vibes?
Hey man I'm just looking up the symbols like they told me
Based on the votes, I can only assume people are still deluding themselves on LLMs capabilities. Is it doing amazing stuff? Yes. But it seems like people still think coding is the ultimate and hardest possible job and so if it can do that it must surely be able to do everything else. My personal experience has show that it still regularly makes up garbage and throws in nonsense sources that do not back up its claims.
Yeah maybe if your topic has 2 decades worth of text material to absorb it will get it mostly right like with coding, but anything that is less common? Complete crap shoot.
Just today I wanted to know if platinum cure silicone will be inhibited by plaster. The first 20 results are all AI spam with 30 pages of fluff and thus unreliable at best, so I asked AI directly. At first it says sulfur and calcium will inhibit the reaction, which is bad because plaster contains those elements. Then it says it will be fine according to X sources. Check the sources, none of them have anything at all to do with curing silicone on plaster, the articles are about using silicone molds to cast plaster. Failure.
Eventually I just had to search youtube videos until I found someone doing it in real life.
I see the same bad, and sometimes catastrophic, takes on things I have a lot of experience in, like agriculture, construction, and mechanics. It is completely worthless for anything mechanical unless you are trying to start something extremely simple from the 40s or earlier, and even then it will still tell you stuff like "clean the carburetor" on an old hot bulb diesel.
> My personal experience has show that it still regularly makes up garbage and throws in nonsense sources that do not back up its claims.
When coding? I feel like it only makes mistakes anywhere near that when my prompting is lazy or stupid. As long as I feed it enough context its really good but it does overfit a lot still.
Some of those were pretty misguided and couldn't be answered yes/no/not sure. For example a post saying that it couldn't produce a convincing illustration misses the point: it can produce coherent and sensible pictures, but they're all too similar. You need novel input and even then there's no guarantee. "Convincing" has way more than one dimension. The problem of output variety was obvious and well known in 2024 when it was posted and is painfully obvious now when everyone is tired of slop, and neither big labs nor other researchers are interested in solving this. Another problem is that the agentic/coding training leaks like crazy, and models became worse in creative department even compared to 2024, despite being more coherent and convincing. So the answer is technically yes (and was yes when it was posted), but practically "it depends", and goalposts stay in the exact same spot they were in 2024.
If for each mistaken prediction there was some mild accountability, like someone shows up and slaps you with a trout, it would improve the site. But it should be added to the terms of service first.
Alternatively, you can bet on your predictions. If you're wrong, you lose money.
you know that's not a bad idea there are a lot of people who are very confident on both sides of the argument. I wonder how many would actually be willing to put their money where their mouth is.
Is it really AGI if it can’t come to my address and slap me with a trout? Clearly AI is all hype /s
My test would be an AI agent has a constantly growing karma HN account that makes comments of various lengths without being detected or banned. Wait...
How was this assembled? From a meta point of view, how much AI was used to curate and highlite the goals; how much was used to assemble the site itself? Or deploy it?
Too many questions. I bailed after about 10, with no idea how many more there were.
This is great and fun even. I'm disenfranchised in America for being part of the DSA so I take the opportunity to vote whenever I can.
A chess scoresheet sometimes contains mistakes but chess players can figure out in many cases what was meant by thinking of what moves make sense and considering the level of play so far. Popular AIs tools fail at that.
Chess is an interesting case. I remember in 2023, GPT 3.5 or something used to be surprisingly good at chess. There was even a "stochastic parrot chess" website [1]. I recall it was playing decently at around a 1800 level. Even as a fairly okay player myself (2100 bullet on lichess), I struggled to beat it. However, modern LLMs are a lot worse at chess. I guess having too much chess data in the training set probably regressed performance on stuff that actually matters, like coding.
[1] parrotchess.com, no longer available. Previous discussions: https://hn.algolia.com/?q=parrotchess.com
> I guess having too much chess data in the training set probably regressed performance
I think the theory is that the LLM having a high chess ELO was a pet project of a researcher that left.
Goalposts do seem to be moving:
Moving or just different goalposts?
It's about how different commenters have defined AGI over the years, so I would say moving the posts.
I made a bet with a guy on HN that the market value of OpenAI + Anthropic would get to at least 2.5T by 2027. I think I'm on track to winning.
https://news.ycombinator.com/item?id=48517353
I also made a bet that API inference margins are greater than 10% for OpenAI and Anthropic
https://news.ycombinator.com/item?id=48500827
I can make another prediction about Agentic Commerce and I think it will get big. Muse + Grok Bot + Dots.
I love it! Kinda wholesome how that heated discussion ended with that bet.
You're very lucky that market value and actual value isn't the same thing.
Not the same thing because there is no "actual value". You should claim that invention.
But there is no market value pre ipo
> API inference margins are greater than 10% for OpenAI and Anthropic
How do you measure that?
With Amodei's special accounting ofc
Which is?
Voting on this is ridiculous. Obviously we should have AI decide which AI challenges have been met.
Not sure how questions are spread among people but so far all of mine have been “no” barring a few from before 2022.
To be fair, none of them have actually been met. Mostly what’s stopping them is the “reliably” part.
All I learned from this is that 40% of hackernews are AI haters which maps pretty well from the overtly negative sentiment on it constantly.
Quite a few of the challenges revolve around asking for LLMs to complete tasks reliably and aren't about whether an instance of an LLM completing the task exists. Quite a few of the goalposts are consequently completely changed without the surrounding context, are not the same as what the HN commenter requested and hence seem disingenuous to me.
Well I said that before AI will soon make the pcb and electronics just like code, it seems some hw engineers didn’t like it, months later there are few products about the same idea :)