Russell
Yes, and so LLMs will evolve into 1) specialized agents (already happening) 2) increasing association between increasingly abstract concepts, 3) competing predictions 4) wayfinding between predictions, 5) auditing predictions for possibility of acting upon them.
Source date (UTC): 2024-07-08 19:07:33 UTC
Original post: https://twitter.com/i/web/status/1810390280912384459
Reply addressees: @RussellJohnston @RolandBasilides
Replying to: https://twitter.com/i/web/status/1810388770660720703
IN REPLY TO:
Unknown author
The reason we require reason so to speak exists because our auto-associative memory (intuition) proposes too uncertain or too many competing (undecidable) possibilities to choose from. Reason then acts as a wayfinding process (navigation, maze-solving) in order to get to some one of those possibilities. And this process iteratively continues until some possibility is chosen as possible. 😉
Original post: https://x.com/i/web/status/1810388770660720703
Leave a Reply