"But.. okay so maybe AGI would be 1000x smarter than us and have better things to do than care for apes. Maybe it wouldn't use earth's resources and would just go into the galaxy - and let us have earth, right? Maybe it'll all be okay, right?" pic.twitter.com/6jBm4bCnNz

This is a reminder that: 1) These models are predictors, not actors. 2) The brain is orders of magnitude more data efficient. 3) AI can't yet drive a car or even play Mario 64. Impressive and useful, but still missing pieces for silicon life (AGI). It's not just scaling compute x.com/stephenbalaban…

Whatever your opinion, I do think that Yoshua Bengio and loads of other big names in AI calling for a pause should be front-page news futureoflife.org/open-letter/pa… pic.twitter.com/WQi7gV9kWQ

Von Neumann on existential risk: Progress is chaotic so investments in planning are much less useful than investments in better control systems. Don't build Maginot lines, develop better command structures. Shorten your OODA loop. Response speed is everything. pic.twitter.com/lnS8s1VWT5