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.

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.

Technology gives and technology 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.