3 ms·
He's explicitly not an AI doomer. He says that he disagrees with Yudkowsky and thinks superintelligence will be fine if we're careful. Curiously he never submit
by hackerlight 3y ago
He's explicitly not an AI doomer. He says that he disagrees with Yudkowsky and thinks superintelligence will be fine if we're careful. Curiously he never submits his reasoning for critique, it's always just an assertion given in softball interviews.
- Vecr 3y agoIf you compare to Yudkowsky, not many people are. Nick Land? Gwern? I think of it, if someone thought there was a 4% chance of a comet killing every single person on Earth before the year 2100, it would be very hard to claim they were an optimist, though that's lower than the p(doom) number "AI optimists" give. It's only relatively that you can say people like Hinton, Christiano, Altman, and Amodei are not doomers.
- hackerlight 3y agoHinton I would say is more in the doomer camp. He is really concerned and says he has no idea what to do aside from raise the alarm for more people to work in AI risk. Christiano puts AI takeover at something like 20 percent chance if I remember correctly. Sam might have similar probabilities but has never verbalized it. It's more a "I disagree with Eliezer but trust me bro, I think we can get it right" attitude in his public facing statements so far. He is more towards dooomer than Zuck, though. At least he recognizes there's a difficult problem there.
- goatlover 3y agoBeing careful means massively funding AI development? It looks a lot to most people like a race to AGI, whatever the consequences.