Idea: write a blogpost about why research is a distribution so it seems a bunch of EAs often forget that this framework of “useful”/“not useful” work is (haha) not useful and we should be looking at EVs. EAs know that but often acts as if they don’t. (A.k.a: empirical people still seem to criticize theory saying it’s useless when actually in the face of short timelines, I don’t really see how some random specific circuit is going to save us from ASI)
Idea: write a blogpost about why research is a distribution so it seems a bunch of EAs often forget that this framework of “useful”/“not useful” work is (haha) not useful and we should be looking at EVs. EAs know that but often acts as if they don’t. (A.k.a: empirical people still seem to criticize theory saying it’s useless when actually in the face of short timelines, I don’t really see how some random specific circuit is going to save us from ASI)