Link post
Saw this on Hacker News; there’s Hacker News discussion here.
Formal methods seem very relevant to AI safety, and I haven’t seen much discussion of them on Less Wrong.
Why don’t people use formal methods?
Link post
Saw this on Hacker News; there’s Hacker News discussion here.
Formal methods seem very relevant to AI safety, and I haven’t seen much discussion of them on Less Wrong.