2026-09-04 06:23 UTC
Marcus is not anti-AI. He is anti Scam Altman and Dario - or more specifically anti-llm. He relentlessly witters on about how “well aktually if you did it my way it would be good AI…” he is still an AI maximalist that wants to bring about AGI. He is one of the most frustrating people whose work I’ve read because he is soooo close to making a good point the majority of the time and then spoils it by saying “but guise we should totally still do AGI and we should do it my way”.
Replies (2)
-
@fnix@awful.systems 2026-09-04 06:48
Seems like he’s very heavily professionally invested in the symbolic paradigm and given machine learning completely dominates AI at the moment he’s taking the piss at them whenever he can. Given the fundamental limitations of ML he’s probably right that any sincere attempt to improve on them requires a more symbolic approach, but talking specifically about AGI is getting into woo territory.
-
@rhamph@awful.systems 2026-09-04 09:03
I suppose my error here is taking him at his word that he could support a ‘pause’, when such a statement could very well just be a concession to appear moderate and reasonable.