Which leads to reminder of a common failure on these topics: use of “believe” in a confusing way. For many mind-killing topics, don’t focus on epistemology—it’s really not easy to get enough agreement on terms that a useful discussion can be had. Instead, focus on predictions.
Not “why do you believe X”, but “what do you predict will happen based on X”.
I think there’s at least half a dozen of Eliezer’s old founding posts that make explicit reference to this concept: real belief has to include likelihood functions.
Which leads to reminder of a common failure on these topics: use of “believe” in a confusing way. For many mind-killing topics, don’t focus on epistemology—it’s really not easy to get enough agreement on terms that a useful discussion can be had. Instead, focus on predictions.
Not “why do you believe X”, but “what do you predict will happen based on X”.
This sounds like an excellent way of cutting through mindkill. If there’s not a post on it somewhere, I think there should be.
[ETA: Bit of a brain fart, sorry. This is making arguments pay rent in the same sense that one should make beliefs pay rent.]
I think there’s at least half a dozen of Eliezer’s old founding posts that make explicit reference to this concept: real belief has to include likelihood functions.