Maybe it’s somewhat in bad taste to propose a project I am involved in, but I think that Max Harm’s and Seth Herd’s ideas on Corrigibility / DWIMAC need support. Ideally, in my eyes, an org focused specifically on it.
See Corrigibility as Singular Target series for details.
Maybe it’s somewhat in bad taste to propose a project I am involved in, but I think that Max Harm’s and Seth Herd’s ideas on Corrigibility / DWIMAC need support. Ideally, in my eyes, an org focused specifically on it.
See Corrigibility as Singular Target series for details.