[–] 1 point 1 month ago (2 children)

A few reasons:

  1. The evidence so far seems to be that AI capability scales with compute, and compute continues to advance. Including with crazy AI hardware like Cerebras and Taalas. There are still a ton of AI hardware startups that have yet to tape out.
  2. There's an enormous amount of research into algorithmic and infrastructure improvements, which also show no signs of slowing down.
  3. Empirically, it has been getting better continuously for several years now. There appears to be no reason to suspect any blockers.

In other words: why not?

To steel man, the main reasons I would counter with are:

  1. Moore's law is pretty much dead... Except that we still have a ton of headroom with custom AI-focused chips. E.g. I know of at least two startups working on DRAM stacked on the compute chip instead of next to it, which gives dramatically higher memory bandwidth.
  2. There are still a couple of architectural issues, e.g. context length is limited, AI can't really remember/learn things on-line, hallucination is not solved. But... in practice those problems seem to not matter too much and also hallucinations seem less common the more capable the model is, so maybe it doesn't need an architectural solution at all.
  • source
  • parent
  • context
  • [–] 11 points 1 month ago* (4 children)

    I agree - Structs are kind of a nicer solution here. Especially because it's pretty common to pass a whole set of options through the call stack.

    In Python land you end up with functions with 30 arguments and when they want to pass them anywhere they give up and use kwargs which are completely awful and should almost never be used.

  • source
  • [–] 1 point 1 month ago (1 child)

    Of course it does. You said he has a higher capacity than AI and he just needs to learn to get there. Well that's no help if by the time he does AI has advanced even further.

    learned helplessness

    No, I just accept reality. You can bury your head in the sand if you like.

  • source
  • parent
  • context
  • [–] 2 points 1 month ago (3 children)

    You said

    It’s better than you right now, but every mid-level tech worker will tell you it sucks. You have significantly higher capacity than it, you just have to learn to get there and that isn’t easy or quick.

    I was just pointing out that this view completely depends on AI improving slower than he does. I would not put money on that.

  • source
  • parent
  • context
  • view more: ‹ prev next ›