Conversation

Notices

1. Embed this notice
   meeper@mai.waifuism.life's status on Thursday, 30-Jan-2025 04:00:14 JST meeper

   Kinda doomer but I feel its very likely we may get some sort of strong ai by 2027. At least for math and programming. Its really not just hype there are legit reasons to believe it.
   
   Hallucinations may still be issues, perhaps context lengths and even inference time costs, creativity may really not be reduced and I wonder about code performance and security and am sceptical of Dario's "A country of PHDs in a datacenter"
   
   But it really seems to be a real deal

   • dilbert 1 likes this.
   • Embed this notice
     dilbert 1 (sun@shitposter.world)'s status on Thursday, 30-Jan-2025 04:00:55 JST dilbert 1

     in reply to
     @meeper I stil don't think we'll see AGI but regular AI will keep getting better