AI 2040 discussion: AI safety planning workshop

A special edition of our workshop. Everyone in this field carries around a number, or refuses to, and both of those are interesting. We'll spend the session actually interrogating where those probabilities come from: what the estimate is doing for you, which cruxes move it, what would change it, and whether the number is load-bearing for anything you do. Bring your laptop and your priors. One note: the date may move. We're working on something that would make this a considerably better session, and if it comes together we'll shift to accommodate it. Register anyway and we'll keep you posted.