you are viewing a single comment's thread
view the rest of the comments
[–] 11 points 11 months ago* (last edited 11 months ago) (5 children)

LLM's are cool and all, but you can't use them for anything requiring real precision without allocating human work time to validate the output, unless you want to end up on the national news for producing something fraudulent.

And making it so their image generator can generate porn isn't going to change that.

  • source
  • parent
  • hideshow 5 child comments
  • [–] 2 points 11 months ago (2 children)

    Yes, currently AI isn't reliable enough to use instead of a human. All the big AI businesses bet that this will change - either by training with more data or some technological breakthrough.

  • source
  • parent
  • hideshow 2 child comments
  • [–] 6 points 11 months ago

    Could be they're right.

    They tried that with Theranos because Elizabeth Holmes' machine could correctly identify four viruses.

    Presumably LLM's have already trained on the entirety of human knowledge and communication and still produce buggy information, so I'm skeptical that it'll work out the way the VC's expect, but we'll see.

  • source
  • parent