I currently think he seems quite wrong.

For context, I work at Redwood Research and I'm more sympathetic to views similar to Paul Christiano or views commonly discussed at Open Phil.

My guess at the probability of doom is more 30-40%.

This is resolved entirely based on subjective judgement.