>And yet Elon Musk is involved in this project. So are Sam Altman and Peter Thiel. So are a bunch of other people whom I know have read Bostrom, are deeply concerned about AI risk, and are pretty clued-in.

This is precisely what dumbfounded me about the announcement.

>My biggest hope is that as usual they are smarter than I am and know something I don’t.

It's possible that OpenAI might be a play to attain a more accurate picture of what constitutes state-of-the-art in the field, effectively robbing the large tech companies of their advantage—all the while building a robust research organization that could potentially go dark if necessary.

Admittedly, that also sounds like it could be the plot to a Marvel movie. Perhaps a simpler explanation is that the details aren't really hashed out yet, and they're essentially going to figure it out as they go—which would be congruent with the gist of OpenAI's launch interview.