The problem is that epistemology is still its own thing, regardless of what you're doing. If your thing is lifting weights, you need well designed, large n, un

The problem is that epistemology is still its own thing, regardless of what you're doing. If your thing is lifting weights, you need well designed, large n, unbiased studies about that. If your thing is cooking, you need studies about cooking. Metis seems to boil down to interest and culture. Those are certainly things that might give you some sensible priors, but in the end you know things because the mechanics of discovery are what they are. Note that noise is a big issue. Any given bodybuilder can only follow one program at a time, and can only change so often. So someone who happens upon a working concept might never really know why it worked.