Noosphere89 comments on The alignment problem from a deep learning perspective

[Noosphere89]

I’m interested in being paid to do this.

Also, the paper makes fairly minimal assumptions, and even in a slow takeoff like a decade or 2, there’s still severe problems.