The promise of upside grows exactly in lockstep with the danger of downside.
It is absolutely rational to say: “if our AI safety research lab is being hijacked by commercial interests before it’s even commercially viable as a business, then how will we possibly pull the plug on the final descent?”
P.S. neither Yann nor me nor you nor Sam nor Greg know what “real intelligence” is, and we certainly don’t know what superhuman intelligence is just the same way an ant couldn’t possibly know what human intelligence is. It is inconceivable to us — that is the point of intelligence.
Point to the physical laws that indicate AGI or superintelligence is impossible or even maybe-impossible. No such laws exist.
I would like to buy Yann's house cat, please. It will get along well with the rather impressive stochastic parrot I bought from someone else on HN.
> It seems to me that before "urgently figuring out how to control AI systems much smarter than us" we need to have the beginning of a hint of a design for a system smarter than a house cat.
and
> don't confuse the superhuman knowledge accumulation and retrieval abilities of current LLMs with actual intelligence
https://x.com/ylecun/status/1791890883425570823