#nerdreic — Public Fediverse posts
Live and recent posts from across the Fediverse tagged #nerdreic, aggregated by home.social.
-
LLMs Don’t.
Model Collapse Ends AI Hype. George D. Montañez, PhD.
LLMs Don’t Think: They process tokens via statistical patterns, lacking internal states or understanding
LLMs Don’t Reason: They exploit superficial cues and rationalize answers post-hoc, failing at adaptive problem-solving
LLMs Don’t Create: They recycle and degrade existing information, unable to escape the "syntax trap" (manipulating symbols without semantic grounding)
https://yewtu.be/ShusuVq32hc or on the #nerdreich ’s attention farm https://youtu.be/ShusuVq32hc