you are viewing a single comment's thread
view the rest of the comments
[–] 13 points 6 months ago (11 children)

I would like to think if I were a student, that I would use AI to explain things to me that I had too much anxiety to ask. I'd probably to use it to do my homework but to mainly check that I did it right? I'd "like" to think that but I wonder how the temptation would get me to just shortcut it ...?

I'm in software and we have to use it for work. There's some toxic positivity going around with it and everyone is just saying how they don't write code any more and how proud they are of that. They even turn their nose at me because I don't use it as much because I'm lower on the Cursor dashboards than them. I like using AI to explain things to me but I'm mainly a senior so I don't write as much code, I actually spend more time reading and it's far more arduous. We actually have juniors turning into prompt monkeys were they just put everything into AI and if you ask them why, they just put your question into AI and giving it back to you. To be clear, that's been happening to other seniors that work with them. If it was me, I'd rip them apart and tell them I can find a quick way to cut cost in the company if they don't wanna learn anything.

  • source
  • hideshow 11 child comments
  • [–] 12 points 6 months ago (2 children)

    … they just put everything into AI and if you ask them why, they just put your question into AI and giving it back to you.

    This must be what it's like for your planet to be taken over by an alien hive mind.

  • source
  • parent
  • hideshow 2 child comments
  • [–] 10 points 6 months ago* (1 child)

    I refuse to believe that the Pluribus series from Vince Gilligan is not actually about AI. There are so many scenes in it that could be adopted to people overusing LLMs.

  • source
  • parent
  • hideshow 1 child comment
  • [–] 1 point 6 months ago (7 children)

    I'd probably to use it to do my homework but to mainly check that I did it right.

    This is where I’ve found it useful. Or I’ll say to not give me the answer but to show me the method I should follow for a problem I don’t know how to solve.

  • source
  • parent
  • hideshow 7 child comments
  • [–] 4 points 6 months ago (6 children)

    The problem is that this doesn't work if you don't know the subject matter already. You can't tell the facts from the inevitable hallucinations.
    A possible mitigation might be to ask the same question in different phrasings multiple times, and ideally ask different models, to distill the truth that way. But I'm not sure you still save any time, that might actually take longer than just learning the old-fashioned way. Plus you don't learn how to learn, so it's a net negative.

  • source
  • parent
  • hideshow 6 child comments
  • [–] 1 point 6 months ago (1 child)

    ask the same question in different phrasings multiple times

    When I've tried that, I often get different answers.

    Or when I try testing the answer with followup questions etc, it eventually devolves to nonsense.

  • source
  • parent
  • hideshow 1 child comment
  • [–] 1 point 6 months ago* (last edited 6 months ago)

    That's why I called it a mitigation, not a solution. Also n may need to be rather big for this approach to work and it might only have a chance of succeeding if the context up to that point is as similar as you can make it.

  • source
  • parent
  • [–] -1 points 6 months ago (3 children)

    What if you are not capable of learning the material on your own? Then it’s just tough luck.

  • source
  • parent
  • hideshow 3 child comments