It seems like AI safety has gained a lot of broader credibility from being ahead of the curve on cyber, so there's definitely precedent there. I guess the concern is that, possibly (or even probably), x-risk is a near-term risk, in which case focus on other stuff needs to be complementary instead of detracting. Definitely a lot of food for thought.
I also think there's some overlap between burning boats and bridges; moves that make sense coalition wise in short timelines e.g. playing to the party in power now is smart if we get takeoff in a year or two but polarizes the issue after that. Or China, who probably doesn't have the chips to compete in short timelines world but who we would need to coordinate with in slower ones. Unfortunately a common dynamic that very different actions are optimal depending on timelines. Yet another reason for transparency.
It seems like AI safety has gained a lot of broader credibility from being ahead of the curve on cyber, so there's definitely precedent there. I guess the concern is that, possibly (or even probably), x-risk is a near-term risk, in which case focus on other stuff needs to be complementary instead of detracting. Definitely a lot of food for thought.
Thanks for the kind words!
I also think there's some overlap between burning boats and bridges; moves that make sense coalition wise in short timelines e.g. playing to the party in power now is smart if we get takeoff in a year or two but polarizes the issue after that. Or China, who probably doesn't have the chips to compete in short timelines world but who we would need to coordinate with in slower ones. Unfortunately a common dynamic that very different actions are optimal depending on timelines. Yet another reason for transparency.