I mean I think he can write his own ticket. If he said "AGI is possible but not with autoregressive approaches" he could still get funding. People want to get behind whatever he is gonna work on. But a certain amount of hype about his work is needed for funding, yes.
Kinda, as long as it’s cool. If he said ‘this is all just plausible text generation’, I think you’d find his options severely limited compared to the alternatives.
