I definitely agree with this. A lot of my criticisms of the general kinds of things that get discussed around here (which I used to voice more often on OB) disappeared when I saw the sorts of problems they were being applied to. The “planning model of rationality” works remarkably well when applied to problems where you get to plan. My initial criticism of Bayesian methods for making decisions under uncertainty was that it doesn’t work very well for most of our decisions, things like “Which path should I take across the room to retrieve my beer?” People just can’t do the relevant math that quickly, and you run into Dennett’s frame problem right away. But when you’re deciding whether it would be a good idea to donate to Givewell, you can do the whole problem at a chalkboard over the course of a few days if you have to.
I definitely agree with this. A lot of my criticisms of the general kinds of things that get discussed around here (which I used to voice more often on OB) disappeared when I saw the sorts of problems they were being applied to. The “planning model of rationality” works remarkably well when applied to problems where you get to plan. My initial criticism of Bayesian methods for making decisions under uncertainty was that it doesn’t work very well for most of our decisions, things like “Which path should I take across the room to retrieve my beer?” People just can’t do the relevant math that quickly, and you run into Dennett’s frame problem right away. But when you’re deciding whether it would be a good idea to donate to Givewell, you can do the whole problem at a chalkboard over the course of a few days if you have to.