If Elon Musk does something as a result of his AI angst by 2025, will Eliezer Yudkowsky judge it to be a positive or neutral initiative (as opposed to negative)?

closes Dec 31, 2025

23%

chance

ALL

Context.

1. https://manifold.markets/Writer/will-elon-musk-do-something-as-a-re?r=V3JpdGVy

2. https://twitter.com/elonmusk/status/1629901954234105857

Feb 26, 8:56pm: ~~If Elon Musk does something as a result of his AI angst by 2025, will Eliezer Yudkowsky judge it to be a dignified or neutral initiative (as opposed to negative/undignified)?~~ → If Elon Musk does something as a result of his AI angst by 2025, will Eliezer Yudkowsky judge it to be a positive or neutral initiative (as opposed to negative)?

Sort by:

Garrett Bakeris predicting NO at 21%

@Gurkenglas Why such confidence in Elon?

Gurkenglasis predicting YES at 21%

@GarrettBaker Resolution criteria unclear, plus a good chance of nothingburger.

Writeris predicting NO at 20%

From this Reuters article. Elon Musk: "I'm a little worried about the AI stuff [...] We need some kind of, like, regulatory authority or something overseeing AI development [...] make sure it's operating in the public interest. It's quite dangerous technology. I fear I may have done some things to accelerate it."

Gigacasting

Robin Green

Elon Musk does not appear to understand AI alignment, although he understands that AI is very dangerous, which is a lot better than very many people in the world. So this could go either way. I guess it depends on if he hires someone who is prepares to educate him, and doesn't fire them for disagreeing with him stridently.

Garrett Bakeris predicting NO at 20%

@RobinGreen Most things I expect him to do would be net negative. The only advantage he has is a shit ton of money, but any apparatus he tries to set up here will surely be goodhearted to oblivion, and because improving capabilities is easier than advancing alignment, and he’ll likely at least be able to distinguish between project speed related to vs unrelated to AI, he will likely inadvertently end up funding capabilities work even if he’s more concerned about existential risk than woke AI.

Garrett Bakeris predicting NO at 20%

@GarrettBaker speed -> s

Garrett Bakerbought Ṁ50 of NO

@GarrettBaker Relatedly: https://www.lesswrong.com/posts/YABJKJ3v97k9sbxwg/what-money-cannot-buy

Robin Greenis predicting NO at 20%

@GarrettBaker I think you meant goodharted. Goodhearted means "well-meaning".

Gurkenglasbought Ṁ200 of YES

Suppose it turns out that for convoluted reasons that nobody saw coming, the initiative technically turns out to leave us better off than if he hadn't done it. Does this resolve by Eliezers judgement upon hearing Elon's decision, or which of any number of variants?

Garrett Bakerbought Ṁ1,000 of NO

https://www.theinformation.com/articles/fighting-woke-ai-musk-recruits-team-to-develop-openai-rival

harfebought Ṁ300 of NO

@GarrettBaker very unlikely that EY would like that

Alex Amadoribought Ṁ5 of YES

@harfe sound's p orthogonal to his concerns imo

Garrett Bakerbought Ṁ700 of NO

@AlexAmadori Starting yet another AGI lab would likely be ranked as Very Bad by Eliezer standards, and if Elon's main concern about OpenAI is that its too woke, then certainly he hasn't learned anything new.

Alex Amadoriis predicting YES at 25%

@GarrettBaker i may be missing something cause I didn't want to sign up to the news letter. i agree he doesn't seem to have learned anything new

Tоmsold Ṁ78 of NO

@GarrettBaker https://twitter.com/ESYudkowsky/status/1630382864486502402

Garrett Bakeris predicting NO at 17%

@tom I saw this, but thanks anyway. The tweet however, made me slightly update in the YES direction.

Eliezer Yudkowsky

Note that "dignified" in this case simply means "positive" - I don't actually think this was a good time for the emotional phrasing. It may confuse potential bettors.