I like to use AI as a sort of brainstorming tool. For example, if I have an idea for a software system, I will describe the general idea and ask for three different possible designs. I will tell it not to generate code but to provide three ideas with pros, cons, and tradeoffs. What it provides will often trigger some new ideas that take me off in a different direction. I often mix and match aspects of the ideas it provides with my own. I keep iterating on this till I think I have a design I like.
I use it to explain and explore things I don't understand and I will have it give me code examples in multiple languages, explain the benefits and differences which each language offers for the task, and ask questions about each. I stick to the ChatGPT free plan and when I hit the usage limit for the top model, I close the window and switch to writing code. by the time my limit resets I will have much better and more focused questions to ask but generally I wait until I have a real reason to go back, bang my head against the problem for awhile.
I have found that it is much better at answering questions if you start with code it wrote instead your own code or someone else's code, so I boil my question down to a simple programming task and start by having it write that code. For example, there were some things I was unsure about with VMs/bytecode interpreters/compilers, so I started my session by asking ChatGPT to write me a simple Forth VM in C and then used that as the jumping off point.
I create specific prompts, as if Gemini were a college professor with a personality, and then I ask questions and research interesting things I find along the way. I also like to note down some topics that I consider important.
Spending time facing a problem and having no idea how to solve it is an incompressible time in the process of learning/understanding something. So I think it could be: 1) forcing yourself to think without any help, 2) asking AI after some time if no solution comes up to you (and there is a certain time constraint to deliver), 3) understanding perfectly the AI produced solution, no matter the time it takes.
Introduce friction. Use it only through an API, and replies are emailed after a day. Have it refuse to answer questions missing a minimal repro case. Let it chasten you if the same question was asked before (within a year or month).
Request quizzes or hints before showing the answer. Have it go through a Socratic dialog style, where you can reply, "Figured it out" to abort.
Kind of funny sounding, but you could ask AI some tips on how to do this...
I think naturally some output from AI will be incorrect and this forces people to question the output constantly, so I'm not sure if there is much actual current risk of such a loss of skills
I believe people will lose their critical thinking skills, when they use AI blindly. May be we should write the code for a requirement, and then ask the AI to review and suggest any optimizations, so that we could probably learn the mistakes we make.
- use it in "learning" mode
- invest time in doing things without AI - to improve the speed of doing things with AI. Training mode (without AI) vs Performance mode (with AI)
In my experience as a consumer, for study and amateur production, the preservation and fortification of critical-thinking is inevitable and prerequisite. Without such, it would be a dark age, with incoherent whispers from the future.
With coding, results are often proportionate to the quality, thoroughness and structure of my prompts. And much debugging.
With intellectual pursuits, any output of interest or previously unknown nature must be verified. I find the farther I push past the safety rails (always ethically, but often harsh) the greater the inclination to hallucinate, appease/placate or evade through reframing, detours and actually quite a number of clever but counterproductive maneuvers.
But I'd say, have no worries. Critical Thinking to the LLM is copper to the cable. It simply doesn't work without it. And because of this, an unintended benefit comes from exercising one's attentiveness and scrutiny. Just keep in mind that a Pandora's Box exists behind scrutiny. Depending on one's level of patience and persistence, this alone is a deep and fascinating pursuit in itself.
The only way AI will kill a skill is the same as anything else would. It's the underlayment, ie indolence, naivety, etc. For the generally aware, it's a force multiplier and for me, new form of existential mirror.
this a logical doom loop.
are you are a fucking punch press operator, or something else, worse?, such as a temporary hire untill the punch press no longer requires you.
my advice would be to pay attention to punch press design and optimisation, and keep your fingers (mind) clear of the works and keep snagging the custom work that cant be fully automated.
I like to use AI as a sort of brainstorming tool. For example, if I have an idea for a software system, I will describe the general idea and ask for three different possible designs. I will tell it not to generate code but to provide three ideas with pros, cons, and tradeoffs. What it provides will often trigger some new ideas that take me off in a different direction. I often mix and match aspects of the ideas it provides with my own. I keep iterating on this till I think I have a design I like.
I use it to explain and explore things I don't understand and I will have it give me code examples in multiple languages, explain the benefits and differences which each language offers for the task, and ask questions about each. I stick to the ChatGPT free plan and when I hit the usage limit for the top model, I close the window and switch to writing code. by the time my limit resets I will have much better and more focused questions to ask but generally I wait until I have a real reason to go back, bang my head against the problem for awhile.
I have found that it is much better at answering questions if you start with code it wrote instead your own code or someone else's code, so I boil my question down to a simple programming task and start by having it write that code. For example, there were some things I was unsure about with VMs/bytecode interpreters/compilers, so I started my session by asking ChatGPT to write me a simple Forth VM in C and then used that as the jumping off point.
I create specific prompts, as if Gemini were a college professor with a personality, and then I ask questions and research interesting things I find along the way. I also like to note down some topics that I consider important.
Spending time facing a problem and having no idea how to solve it is an incompressible time in the process of learning/understanding something. So I think it could be: 1) forcing yourself to think without any help, 2) asking AI after some time if no solution comes up to you (and there is a certain time constraint to deliver), 3) understanding perfectly the AI produced solution, no matter the time it takes.
Introduce friction. Use it only through an API, and replies are emailed after a day. Have it refuse to answer questions missing a minimal repro case. Let it chasten you if the same question was asked before (within a year or month).
Request quizzes or hints before showing the answer. Have it go through a Socratic dialog style, where you can reply, "Figured it out" to abort.
Kind of funny sounding, but you could ask AI some tips on how to do this...
I think naturally some output from AI will be incorrect and this forces people to question the output constantly, so I'm not sure if there is much actual current risk of such a loss of skills
This if you do have some critical thinking to start with, for the less gifted ones LLMs only make things worse.
I believe people will lose their critical thinking skills, when they use AI blindly. May be we should write the code for a requirement, and then ask the AI to review and suggest any optimizations, so that we could probably learn the mistakes we make.
- use it in "learning" mode - invest time in doing things without AI - to improve the speed of doing things with AI. Training mode (without AI) vs Performance mode (with AI)
I paste URLs to docs into Gemini and ask it to generate quizzes.
In my experience as a consumer, for study and amateur production, the preservation and fortification of critical-thinking is inevitable and prerequisite. Without such, it would be a dark age, with incoherent whispers from the future.
With coding, results are often proportionate to the quality, thoroughness and structure of my prompts. And much debugging.
With intellectual pursuits, any output of interest or previously unknown nature must be verified. I find the farther I push past the safety rails (always ethically, but often harsh) the greater the inclination to hallucinate, appease/placate or evade through reframing, detours and actually quite a number of clever but counterproductive maneuvers.
But I'd say, have no worries. Critical Thinking to the LLM is copper to the cable. It simply doesn't work without it. And because of this, an unintended benefit comes from exercising one's attentiveness and scrutiny. Just keep in mind that a Pandora's Box exists behind scrutiny. Depending on one's level of patience and persistence, this alone is a deep and fascinating pursuit in itself.
The only way AI will kill a skill is the same as anything else would. It's the underlayment, ie indolence, naivety, etc. For the generally aware, it's a force multiplier and for me, new form of existential mirror.
Technology gives and technology takes.
How can I sit in a chair all day and not ruin my cardiovascular health?
I think the answer is, you cannot.
this a logical doom loop. are you are a fucking punch press operator, or something else, worse?, such as a temporary hire untill the punch press no longer requires you. my advice would be to pay attention to punch press design and optimisation, and keep your fingers (mind) clear of the works and keep snagging the custom work that cant be fully automated.
John Henry
https://m.youtube.com/watch?v=uYh9MwKQnb8