Regarding the talking point "Why we think there should be a CFAR", what are the best things that could exist instead of a CFAR? Are there any variants that would maintain CFAR's benefits while attracting fewer controversies?
Do you think the AI risk for computer scientists workshops turn anyone off AI risk? How does this compare to the degree to which regular workshops turn people off rationality? Do these workshops target audiences of interest to the AI risk community in general (successful computer scientists in general), or audiences that are of special interest to MIRI (e.g. rationalist pure mathematicians)?
1. What rough fraction of the benefits of CFAR do you think accrue through recruiting for MIRI?
2. What rough fraction of CFAR's revenues come from MIRI?