jsteinhardt comments on Steelmaning AI risk critiques

[jsteinhardt]

For an argument against at least some of MIRI’s technical agenda, see Paul Christiano’s medium post.