Safety
@GaryMarcus Essentially your prediction that LLMs to scale wouldn’t solve key fundamental problems, reasoning, factual reliability, composit…
@GaryMarcus Essentially your prediction that LLMs to scale wouldn’t solve key fundamental problems, reasoning, factual reliability, compositionality, grounded understanding etc are now elephantine chi
@GaryMarcus Essentially your prediction that LLMs to scale wouldn’t solve key fundamental problems, reasoning, factual reliability, compositionality, grounded understanding etc are now elephantine chickens home to roost. It’ll be interesting to see how the various labs try to address it.
Related
- 💯. “@garymarcus has won, others too” And the smart people already know it. AGI requires both symbols and neural networks. Period.
- 👇 “I still maintain that LLMs are really dumb and of limited use. it’s my opinion as a practitioner and professional that the hype being pe…
- We are not getting to the G in Artificial General Intelligence; we are getting to (impressive) advances in particular areas where particular…
- To anyone who read Rebooting AI back in (checks notes) 2019, this is both hilarious and unsuprising. The field has wasted 7 years on an arch…
Source: Gary Marcus (X) | 2026-04-16