#24
#24
@fchollet I remember the interview you did after o3, you were so candid and clear about updating your position. It was incredibly refreshing, and I feel like unfortunately I haven't seen many people handle their own updates with as much grace. Thank you for being intellectually…
@rbhar90 But who ever argued that LLMs would be confined to just a chat window and only allowed the ability to communicate via text? TTC and harnesses are NOT outside the LLM paradigm, especially considering that they are even trained on them. This is just silly.
@GaryMarcus I've been saying the same thing and actually even an LLM will tell you AGI needs more pieces like world models and embodiment, etc. LLM scaling will get better but has a limit/wall coming soon.
@GaryMarcus @fchollet You have the worst takes I've ever seen in AI. You really should stop, else you'll look even more ridiculous.
@GaryMarcus @fchollet And this is why you’re the genius of our age. https://suno.com/s/AZfYkbFvmAiKU3r5