Promoted to curated: I disagree with lots of stuff in this sequence, but I still found it the best case for corrigibility as a concept that I have seen so far. I in-particular liked 2. Corrigibility Intuition as just a giant list of examples of corrigibility, which I feel like did a better job of defining what corrigible behavior should look like than any other definitions I’ve seen.
I also have lots of other thoughts on the details, but I am behind on curation and it’s probably not the right call for me to delay curation to write a giant response essay, though I do think it would be valuable.
Promoted to curated: I disagree with lots of stuff in this sequence, but I still found it the best case for corrigibility as a concept that I have seen so far. I in-particular liked 2. Corrigibility Intuition as just a giant list of examples of corrigibility, which I feel like did a better job of defining what corrigible behavior should look like than any other definitions I’ve seen.
I also have lots of other thoughts on the details, but I am behind on curation and it’s probably not the right call for me to delay curation to write a giant response essay, though I do think it would be valuable.