So what do you do?

I work in AI alignment problems.

See there’s a chance humanity will make an AI and by accident it won’t want to kill all the humans even though we obviously deserve to die.

I’m trying to find a way to guarantee the robot both sees us as an enemy and believes the only solution is total annihilation.

Wow. What’s your method?

Twitter mostly.

Twitter plus various violent dictatorships, but mostly Twitter.

Also is it really fair to tell them to align with us when we don't align with each other?