How do we solve alignment? - Gaurav Yadav

​This week Gaurav is going to give a talk titled 'All the things I'm confused about with extreme power concentration from AI'. He is going to lay out some arguments for why power concentration might get worse with advanced AI systems, why he is somewhat sceptical of this happening, and then invite people to critique his world model. Perhaps 'How Do We Solve Alignment?' has focused largely on technical research so far, but this is a step in a somewhat different direction, broadly taking into account other kinds of things we should be worried about with AI.

It might be worth listening to the following podcast: https://newsletter.forethought.org/p/checks-balances-and-power-concentration and maybe this https://80000hours.org/problem-profiles/extreme-power-concentration/