I think the post is doing three things, all of which I like. First, it documents what it was like for Joe as he made substantial updates about the world. Secondly, it exhibits the rationalist practice of explaining what those updates look like using the framework of probabilities, and considering what sorts of updates a rational agent would make in his position, and contrasted that with a helpful explicit model of how a human being would make updates (e.g. using its guts). And third it’s a serious and sincere account of something that I care about and Joe cares about. I felt reading this post that I was finally sharing the same mental universe as the author (and likely other people reading the post).
There’s lots of more specific things to say that I don’t have the time to, but I’ll say that the paragraph that explains you can’t have a >50% chance of your credence later doubling (or a >10% chance of 10x-ing your credence) struck me immediately as a go-to tool I want to add to my mental toolkit for figuring out what probability I assign to a given statement.
Curated! I loved a lot of things about this post.
I think the post is doing three things, all of which I like. First, it documents what it was like for Joe as he made substantial updates about the world. Secondly, it exhibits the rationalist practice of explaining what those updates look like using the framework of probabilities, and considering what sorts of updates a rational agent would make in his position, and contrasted that with a helpful explicit model of how a human being would make updates (e.g. using its guts). And third it’s a serious and sincere account of something that I care about and Joe cares about. I felt reading this post that I was finally sharing the same mental universe as the author (and likely other people reading the post).
There’s lots of more specific things to say that I don’t have the time to, but I’ll say that the paragraph that explains you can’t have a >50% chance of your credence later doubling (or a >10% chance of 10x-ing your credence) struck me immediately as a go-to tool I want to add to my mental toolkit for figuring out what probability I assign to a given statement.