all 17 comments

sorted by: hot top controversial new old
[–] 50 points 1 month ago (5 children)

in hindsight, believing robots would be compulsively honest was a bit optimistic

  • source
  • hideshow 5 child comments
  • [–] 27 points 1 month ago (4 children)

    They're not really being dishonest, per se. They're just bullshitters spouting bullshit. They don't care about true or false. Because the truth does not appear in the objective functions they are optimizing.

  • source
  • parent
  • hideshow 4 child comments
  • [–] 12 points 1 month ago (2 children)

    Are we talking about AI or politicians here?

  • source
  • parent
  • hideshow 2 child comments
  • [–] 3 points 1 month ago

    i think that gives them too much credit, even. They don't know what is true or false, LLMs can't even really properly do logic in the way that normal programs can, they just have to emulate logic via really fancy statistics and the fact that the things they're trained on contain responses to logic.

  • source
  • parent
  • [–] 48 points 1 month ago (2 children)
  • [–] 26 points 1 month ago (2 children)

    If you need anything, just let me know.

  • source
  • hideshow 2 child comments
  • [–] 20 points 1 month ago (2 children)

    Let me break it down for you

  • source
  • hideshow 2 child comments
  • [–] 12 points 1 month ago

    Good catch!

  • source
  • [–] 4 points 1 month ago

    This is the way

  • source