all 29 comments

sorted by: hot top controversial new old
[–] 43 points 6 months ago (16 children)

...training them to be skeptical and not blindly trust what comes out of the machine...

This is always what I don't understand about using ai in it's current form. If you can't know if it's right or wrong, and have to double check it, why use it in the first place? Would it not be more efficient and easier to just use the couple of petaflops you have in your own head to solve the problem or write that email?

I think then, that it is more of a novelty that has yet to ware off for some people and is conisistently buoyed by the ceos that push it.

  • source
  • hideshow 16 child comments
  • [–] 2 points 6 months ago* (last edited 6 months ago)

    If you can't know if it's right or wrong, and have to double check it, why use it in the first place?

    "If you can't trust that a friend solved a sudoku puzzle for you without checking it first, why even bother?"

    The obvious answer being that it's much easier to check the solution to a sudoku puzzle than it is to solve it yourself. If you have reasonable means to check compared to going out and starting from scratch, then even a modest enough rate of correct answers can save a ton of time. LLMs don't have that for me, but that's also because I've been doing research as a hobby for 10 years.

    If you know anything about computation theory, there's an entire class of problems for which checking a solution is (relatively) trivial but finding a correct one is highly non-trivial.

  • source
  • parent
  • [–] 1 point 6 months ago* (last edited 6 months ago) (2 children)

    It's easier to copywriteedit an email that to write it from scratch.

    Edit: I meant copyedit, not copywrite

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

    If you can’t know if it’s right or wrong, and have to double check it, why use it in the first place?

    Me and my partner alternate doing the cooking. She doesn't know if I'm going to make a mistake and serve her something she doesn't like (it has happened). Does that mean she's better off doing all the cooking herself?

    "If it's not perfect, it's useless" is a fallacy. So the question is, how good does it have to be to be useful? That depends on the task, and especially on the cost (however you measure it - dollars or hours or whatever) of verifying whether the result is good compare to the cost of a person doing the task.

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

    When you cook well, you can eat the food.

    When the bot says something, you always need to look up if it's correct. That's the 'cook a new meal from scratch' bit, not the 'taste it' bit.

    You need to look things up every time, or do the taste test by asking if the bot's answer 'smells true' (which is tempting, but a bad idea).

  • source
  • parent
  • hideshow 1 child comment
  • [–] 3 points 6 months ago (5 children)

    "Food I don't like" as an output isn't really comparible to "information that is factually incorrect."

  • source
  • parent
  • hideshow 5 child comments
  • [–] 0 points 6 months ago (4 children)

    It's comparable because it's a negative outcome that may cost something (cooking a new meal, ordering a takeaway) to fix, but can be checked quite easily. Information that is factually incorrect has a negative outcome as well, and can also be checked quite easily - but the negative outcome, and the ease of checking, varies vastly across the space of all possible information.

    I am encouraging you to think about situations where the negative outcome is not that bad, and the ease of checking quite high. Does that make using AI more practical?

  • source
  • parent
  • hideshow 4 child comments
  • [–] 3 points 6 months ago (3 children)

    Its subjective vs objective. They're not really comparable at all.

  • source
  • parent
  • hideshow 3 child comments
  • [–] 0 points 6 months ago (2 children)

    The objective reality of an AI hallucination being wrong is not what's important though; what is important is the effect it has on people, which will in part be subjective.

    Nothing prevents you from comparing harms and ease of checking.

  • source
  • parent
  • hideshow 2 child comments
  • [–] -2 points 6 months ago

    The petaflops sometimes... flop. The only use case I personally have for llms, and they are brilliant at that, is when a word just won't come to mind—I can give it a precise description of it but my brain refuses to produce the word, in English nor Spanish.

  • source
  • parent
  • [–] 21 points 6 months ago (4 children)
  • [–] 5 points 6 months ago (2 children)

    The Ubuntu users who cared have already left.

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

    "Ubuntu users lay out a plan for switching distros" ... got it.

  • source
  • hideshow 2 child comments
  • [–] 4 points 6 months ago

    the blog post phoronix linked which the verge references https://discourse.ubuntu.com/t/the-future-of-ai-in-ubuntu/81130

  • source
  • [–] 6 points 6 months ago

    Ubuntu will do well if they start with an ai package that can be installed, and let the users slowly test and adopt it, anything else than that and people will just switch distro.

  • source
  • [–] 2 points 5 months ago

    Or, novel idea, maybe NOT do that?

  • source
  • [–] 1 point 5 months ago

    Hello Debian

  • source