If you think 30% is an accurate prediction, there's some price correction in this market: https://manifold.markets/nsokolsky/in-which-year-would-zvi-confirm-tha

@wrhall most likely outcome is that there’s an AI winter but Zvi doesn’t notice

@JesWolfe no new capabilities for half a year and he didn't notice? Pretty much doubt it.

@ICRainbow it's fair to say "criteria are fairly different, hence the discrepancy"

I'm not sure I find the criteria more compelling here (eg if Facebook or Microsoft reduces their purchases of new GPUs, but they keep releasing new models, is that an ai winter? Zvi might not think so, but this q might)

This has been true for most of a decade, but if you took "current ml state of the art" and spent the next two decades finding everywhere that would benefit from it, you would provide benefits to the economy even without new research (this was a Jeff Dean quote from >6 years ago)

That's still true and the state of the art has improved. But some might find new research necessary to avoid a winter

Boosted

I feel like the resolution condition is less interesting than the question I would want answered. I could see all research fronts stalling out, but annual global AI investment still increasing purely on applications.

fuckin lol

I give it like 1% probability but I'm not gonna bet for resolution objectiveness reasons.

Actually wait, a drawdown would not be 50% drop relative to now, but rather 50% drop relative to the peak, right? So maybe it could actually happen.

@tailcalled Yeah that's right, it'd be a 50% drop relative to any previous peak.

@tailcalled Also would be great if you knew a good source for "annual global AI investment" because I couldn't really find one lol. I think Crunchbase might have ~that data, but you need to pay for it.

@erwald If I google "crunchbase AI investment", then I can see that people regularly post diagrams from crunchbase on AI investment, so I think we can just rely on that for resolution. Otherwise I guess we'll have to figure something out at that time. I can probably afford to pay if all else fails.

AI investment could easily drop by 50% in a big financial crisis or us v chin war or something

@jacksonpolack No that would increase it because we would need it for military operations as well. In the words of Vladimir Putin "Whoever becomes the leader in this sphere will become the ruler of the world."

Investment isn't a sense of "how much we need X", it's literally when eg a VC gets 10% of a company in exchange for $200M. The crunch to the economy and trade from a war would cause investment in the latter sense to dry up. The military would up spending on AI, but that'd be dwarfed by less investment in future tech.

@jacksonpolack Perhaps if it were literally world war III. I don't think that will likely happen by 2030 though, I think tensions will increase, which WOULD increase investment in AI.

But the military gets a lot of money, even then, I'm not sure it would be dwarfed by the lower investment in future tech like you claim.