I don’t mean to be difficult. I’m neurodivergent

  • 0 Posts
  • 19 Comments
Joined 3 months ago
cake
Cake day: March 26th, 2025

help-circle






  • I use it almost every day, and most of those days, it says something incorrect. That’s okay for my purposes because I can plainly see that it’s incorrect. I’m using it as an assistant, and I’m the one who is deciding whether to take its not-always-reliable advice.

    I would HARDLY contemplate turning it loose to handle things unsupervised. It just isn’t that good, or even close.

    These CEOs and others who are trying to replace CSRs are caught up in the hype from Eric Schmidt and others who proclaim “no programmers in 4 months” and similar. Well, he said that about 2 months ago and, yeah, nah. Nah.

    If that day comes, it won’t be soon, and it’ll take many, many small, hard-won advancements. As they say, there is no free lunch in AI.



  • If you don’t play chess, the Atari is probably going to beat you as well.

    LLMs are only good at things to the extent that they have been well-trained in the relevant areas. Not just learning to predict text string sequences, but reinforcement learning after that, where a human or some other agent says “this answer is better than that one” enough times in enough of the right contexts. It mimics the way humans learn, which is through repeated and diverse exposure.

    If they set up a system to train it against some chess program, or (much simpler) simply gave it a tool call, it would do much better. Tool calling already exists and would be by far the easiest way.

    It could also be instructed to write a chess solver program and then run it, at which point it would be on par with the Atari, but it wouldn’t compete well with a serious chess solver.