> "Safety issues" are to him more like, "the AI will kill everyone to fill its objectives"

Yes, and he was way ahead of the curve here, since similar positions got a lot more mainstream in the past years and months. E.g. two of three AI Turing award winners (Hinton and Bengio) now say that superintelligent AI poses a serious extinction risk. OpenAI has also identified the alignment problem as a major issue. Even former AI skeptics like Douglas Hofstadter now broadly agree with this assessment. Yudkowsky's opinion was quite prescient.

So in other words, people who've spent their entire lives thinking and talking about AI praise other people who also have spent their entire lives thinking and talking about AI.

This is a bubble conversation that makes no sense to people outside of that bubble, and for good reason; it doesn't matter outside of that bubble.

No, AI experts which formerly were skeptical of an outsider's (Yudkowsky's) opinion, now changed their mind about it.

Wait, so you deny the fact that Hinton and Bengio have worked on AI their entire careers?

No, I don't, why do you think that?

Because you said “no”, and I only said one thing. Do you take that back?

I was disagreeing with it being about experts "praising" each other.

No you weren't, you were choosing to focus on a different category to which these people belonged in an ultimately non sequitur comment.

That's so embarrassingly obtuse I refuse to believe you aren't aware and are just being contrarian.

Not obtuse at all, just catching rhetorical trickery.

None of those people you mentioned share Yudkowsky’s conclusion that air strikes against data centers are warranted.

Do you have a citation for that?

I know how to find the statements by Hinton and Benigo that superintelligent AI poses a serious extinction risk, but I can't find any statements by Hinton and Benigo--or anyone who is taking the risk seriously--suggesting that air strikes would not be warranted, so any clues on where to look would be appreciated.