Hzn comments on Thane Ruthenis’s Shortform

[Hzn]

I think super human AI is inherently very easy. I can’t comment on the reliability of those accounts. But the technical claims seem plausible.