all 5 comments

sorted by: hot top controversial new old
[–] 27 points 5 months ago (3 children)

LLMs are not thinking machines. They are word association probability engines. Most mentions of nukes would be in fiction...and the only reason to mention a nuke in fiction is for the nuke to be used or threatened to be used.

So, yeah. Maybe let's not rely on glorified autocorrect to handle our nuclear arsenals...

  • source
  • hideshow 3 child comments
  • [–] 1 point 5 months ago* (last edited 5 months ago) (2 children)

    I think this YouTube short is a great example of how the AIs are not using reasoning or logic to draw conclusions. https://youtube.com/shorts/bsl46vGpMNU

  • source
  • parent
  • hideshow 2 child comments
  • [–] 5 points 5 months ago

    I think there was a movie on this hehe

  • source