@DavidDuvenaud
Machine learning prof @UofT. Former team lead at Anthropic. Working on generative models, inference, & latent structure.
@gabriel_weil @Jess_Riedel @ESRogs I agree. I think I actually disagreed with you back at "Given the cost of that time, you need to spend a lot on safety for that to be worth it.". The arguments for astronomical waste cut both ways if you think there's a substantial chance of being wiped out.
There’s no textbook on post-AGI political economy. But there are recorded talks from the 2nd post-AGI workshop! All are now online at post-agi.org/talks/ A 🧵 of new highlights:
First, @richardmcngo‘s “favorite talk he’s ever given”: on mechanisms that emergently incentivize the deliberate destruction of value: youtube.com/watch?v=6sf3zQ…
@davidmanheim @deanwball @eligerhard Agree, but even if we enforce a "human actually in the loop", presumably many people will be willing to sell their notional oversight and liability for cheap.
Some new speakers at our post-AGI workshop: - @paulfchristiano on "A typical, reasonably-good future" - @erikbryn on "Post-AGI Centralized Coordination" (i.e. nber.org/system/files/c…) - Sean Lavelle of the U.S. Navy on "AGI and Military Strategic Instability" x.com/DavidDuvenaud/…
