What are supposedly my timelines?

@EliezerYudkowsky I think people perceive you to have very fast timelines but I'm not sure you've ever outlined them in detail.

@EliezerYudkowsky

@SemioticRivalry I remember him saying that if you have children now, there's a good chance you'd still see them live to kindergarten so he probably has a low P(doom) for the next five years at least.

@EliezerYudkowsky You are at +44% on "artificial superintelligence exist by 2030?". Which I think is a quite fast timeline.
(I am at ~2% for it by 2030, and ~25% by 2100)
Otherwise, I think the majority of your arguments are sound.

@EliezerYudkowsky "Your children may get to start kindergarten" strongly suggests 5-10 years.

@EliezerYudkowsky As of December 2021:

I consider naming particular years to be a cognitively harmful sort of activity; I have refrained from trying to translate my brain's native intuitions about this into probabilities, for fear that my verbalized probabilities will be stupider than my intuitions if I try to put weight on them. What feelings I do have, I worry may be unwise to voice; AGI timelines, in my own experience, are not great for one's mental health, and I worry that other people seem to have weaker immune systems than even my own. But I suppose I cannot but acknowledge that my outward behavior seems to reveal a distribution whose median seems to fall well before 2050.

https://www.lesswrong.com/posts/ax695frGJEzGxFBK4/biology-inspired-agi-timelines-the-trick-that-never-works

As of April 2023:

When the insider conversation is about the grief of seeing your daughter lose her first tooth, and thinking she’s not going to get a chance to grow up, I believe we are past the point of playing political chess about a six-month moratorium.

https://www.lesswrong.com/posts/oM9pEezyCb4dCsuKq/pausing-ai-developments-isn-t-enough-we-need-to-shut-it-all-1

And as quoted earlier:

@EliezerYudkowsky You bought YES on "will AI wipe out humanity by 2030" up to 40%, so that shows a minimum probability of 40% chance of AI capable of wiping out humanity by 2030, in 6.5 years.

That timeline is wrong, along with your opinions about AI risk.

if my p(doom) is 30% and his p(doom) is 99.99%, would I say he's 'right'? well, I am closer to him than probably 99.99% of people are, but I am also quite far away in a numerical or log sense. So it's hard to answer.

Also, my understanding is that he doesn't have a very strong position on timelines.

@SemioticRivalry Hm, fair point. I think in that case whether you think he's "right" depends on the vibes. (Is his p(doom) seriously 99.99%????)

He's posted "if we're still alive" on tweets talking about 2-4 year time-frames, so he definitely expects very rapid AI development.

@evergreenemily I don't think he's ever given out an actual numerical p(doom), but I know he's said "probability approaching 1" before.


I think on timelines he has very wide error bars, and that tweet was just a one-off line. When talking with Hotz for example he basically refused to give a timeframe.

He probably won't respond, but might as well @EliezerYudkowsky and ask for specifics

@SemioticRivalry He bought "will AI wipe out humanity by 2030" up to 40%, which is a pretty strong opinion on timelines.

@DavidBolin Looking at

/MartinRandall/will-ai-wipe-out-humanity-before-th-d8733b2114a8

I see bets up to 25% and also later bets down to 16%.

I know he also bet on

/EliezerYudkowsky/will-ai-wipe-out-humanity-by-2030-r but I don't think play money bets on non-predictive markets are evidence of strong opinions.