Capabilities being more jagged reduces p(doom), less jagged increases it.

Ceteris paribus, perhaps, but I think the more important factor is that more jagged capabilities imply a faster timeline. In order to be an existential threat, an AI only needs to have one superhuman ability that suffices for victory (be that superpersuasion or hacking or certain kinds of engineering etc), rather than needing to exceed human capabilities across the board.

rot13: Gur snvyfnsr jnf gur bayl cneg gung fhecevfrq zr, naq gur qvrtrgvp rkcynangvba sbe ubj guvf frghc pbhyq unir unccrarq ng nyy jnf phgr.

I may never stop finding it funny the extent to which Trump will seek out the one thing we know definitively is going badly, then and choose that to lie and brag about.

We're not the target audience for those posts. He's telling everyone who's kissing his ass that this is a topic he wants them to lie about as well.

If we were doing a better job catching cheaters presumably people would be doing it less?

Seems trivially easy for an idiot to start an account, lose some games, get frustrated, and start asking a chess bot what to do so that they can rescue a bad position / show that bastard what's what / vicariously enjoy winning. I expect that crowd to be the majority of cheaters, and for them to be essentially inelastic to the probability with which they get caught (the first time).

I believe that Cade knows perfectly well what everyone has been saying for years; he's being disingenuous because the object level doesn't matter to him, and the only important thing is ensuring that these weirdos don't get status. He's never once engaged on simulacrum level 1 with this community.

At long last, Elon Musk has created the Yandere Simulator from the classic short story Don't Create The Yandere Simulator, Among Other Things

On several of these I think you might be confounding the positively correlated variables "has heard of this question" and "is on the autism spectrum", the latter of which is anticorrelated with the kind of behavior you want (intuitive, empathetic listening without needing to ask a bunch of explicit questions). I find it unlikely that asking the question makes people worse at the behavior you want.

Zvi is arguing "X implies Y" here. Zvi happens to believe Y but disbelieve X; however, he is writing to people who think "X and not-Y", in order to nudge them to support Y.

Here X = it is good for the US to build superintelligence fast, before China does, and Y = we should have some diffusion rules making it harder for China to catch up to the USA. 

Zvi believes Z = nobody should be building superintelligence soon, and believes Z implies Y, but it is useful to show that X implies Y as well.

To be fair, I also think it was a correct decision to focus on the Democratic Party when advocating for sane AI notkilleveryoneism policy before the 2024 election, because that was playing to our outs. Trump will do on AI whatever he is personally paid to do, and the accelerationists have more money; anyone sane but unable/unwilling to bribe Trump has no influence whatsoever with him in this administration.

(I went around saying the above before the election, by the way.)

(Sorry for the mindkilling, it's relevant on the object level.)