@jachiam0 @lavish_lerven And I'm not as sure that Google is less constrained than OpenAI, given their willingness to say there's a deep problem that may require slowing down.

OpenAI has been far more accelerationist and anti-safety, especially at a political + leadership level.

@davidmanheim You're right in saying that there's an element of hubris to expect to be able to control a smarter being. But do you want to then completely stop all progress? Never try for AGI / ASI?