...I for one didn’t take the self-criticism stuff particularly seriously, or consider EA to have scored any points by running the contest? And I thought that seemed too obvious to mention, I guess, that of course scoring diligent self-criticism points is harder than that? Zvi had a longer writeup on a similar take.
I have my own sense of what EA is doing all wrong, but it didn’t particularly occur to me to try to write it up for the criticism contest. That was obviously going to be an in-frame thing, and was obviously not going to be able to reward anything out-of-frame; doing that is really hard if you don’t want to just completely throw the contest over to wild randos. Like, the reason MIRI doesn’t run a contest like that is not that our underlying reality contains nothing worth criticizing, but that a contest like that obviously does not work to highlight your real actual problems.
There’s nothing particularly unhealthy about a movement where a contest like that fails to turn up any real issues—the contest is obviously doomed in that sense from the start. The only failure is the naivety required to imagine that a contest like that could possibly work, against a semiefficient background where smart people have already been saying various things and there are already reasons why those critiques have not been accepted. The best you can say about a contest like that is that maybe it’s worth $100K to get better centralized writeups of in-frame criticisms that your community already knows how to accept but hasn’t finished acting on yet.
...I for one didn’t take the self-criticism stuff particularly seriously, or consider EA to have scored any points by running the contest? And I thought that seemed too obvious to mention, I guess, that of course scoring diligent self-criticism points is harder than that? Zvi had a longer writeup on a similar take.
I have my own sense of what EA is doing all wrong, but it didn’t particularly occur to me to try to write it up for the criticism contest. That was obviously going to be an in-frame thing, and was obviously not going to be able to reward anything out-of-frame; doing that is really hard if you don’t want to just completely throw the contest over to wild randos. Like, the reason MIRI doesn’t run a contest like that is not that our underlying reality contains nothing worth criticizing, but that a contest like that obviously does not work to highlight your real actual problems.
There’s nothing particularly unhealthy about a movement where a contest like that fails to turn up any real issues—the contest is obviously doomed in that sense from the start. The only failure is the naivety required to imagine that a contest like that could possibly work, against a semiefficient background where smart people have already been saying various things and there are already reasons why those critiques have not been accepted. The best you can say about a contest like that is that maybe it’s worth $100K to get better centralized writeups of in-frame criticisms that your community already knows how to accept but hasn’t finished acting on yet.
Have you written that up anywhere? Would be interesting to read.