On: Notes on existential risk from artificial superintelligence

[c22]

I suspect multiple groups are already working on this sort of stuff. I don't expect all of them to succeed, but it only takes a few...