Related to ACX five year predictions. I will resolve this based on my impression of the consensus of economists at that time. By "visible break", I mean clearly larger than ordinary year-to-year variation, and widely remarked upon.
Update 2025-19-01 (PST) (AI summary of creator comment): Clarifications:
A direct effect of AI refers to economic changes directly caused by AI itself, and does not include effects resulting from phenomena like an AI industry bubble bursting.
1,000This is asking about a visible break in economic statistics in <3 years from now. Economic statistics typically have several months of lag time for collection and revision. So it seems that the visible break would need to start happening in the real economy by mid 2027 at the latest.
Then there is also the time required for economist consensus to form and become widely remarked upon. It can be hard to determine causation of macro trends in real time. Even if AI is having major effects, separating its impact from other factors (monetary policy, regulatory changes, global events, etc.) may take time.
Given these constraints, for this to resolve YES, we'd basically need to see dramatic AI-driven macroeconomic changes starting in the next 18-24 months that are so clear and pronounced that economists can quickly reach consensus about their cause.
It seems like this market has settled at basically the same level has the AI movie market:
https://manifold.markets/ScottAlexander/in-2028-will-an-ai-be-able-to-gener
I guess the two are somewhat correlated?
@TimothyJohnson5c16 I'm interested by the fact that they're both quite a bit lower than this AGI one: https://manifold.markets/RemNi/will-we-get-agi-before-2028-ff560f9e9346
If an AI can make a movie, does that make it AGI? The markets say maybe.
Given that humans can make movies, it seems to me that an AGI should be able to. The markets say otherwise!
I think this market (the trend line one) should be the lowest. If nothing else, because there's a lag between something happening and being able to see it in a trend line. But also there's lots of ways something can be cool and impressive without it revolutionising the economy immediately.
@Fion Hmm, good point. The creator for that one deleted his account though, so it's going to be up to the mods to decide what counts as AGI.
@elf The latter. My lived experience strongly indicates that the hype surrounding AI since the release of ChatGPT has vastly exceeded the actual progress. I think it's quite similar to the dot-com bubble. It will bear fruit eventually, but not anywhere as soon as (many) people think.
I don't think there has been any case where one technology alone caused a visible break in GDP, GDP per capita, unemployment, or productivity since the industrial revolution. Most likely AI technology is already priced into current growth projections.
Pretty easy no. Expect to hit big here.
@AdamK I'm one of the biggest holders already. But just to spite you and because I'm 4 beers deep watching CFB playoffs, I will
@JaundicedBaboon Thanks! This is one of the best-priced markets for my marginal mana, so if people want to keep limit orders up, I'll fill them as I get more mana
In my modal scenario, I see the improvements in AI until 2028 as basically a continuation of improvements in computer technologies over the last couple of decades. It seems there would have to be some sort of phase transition to cause a visible break in the trend line of one of these very major trends. As Robert Solow said, "You can see the computer age everywhere but in the productivity statistics."
Worth noting that a lot of these are lagging indicators. AI might produce unprecedented scientific discoveries in 2027, and the stock market might ~double, with little-to-no effect on GDP in 2028.
Does this resolve YES if there's a huge AI bubble that bursts and brings the whole economy down? Say on the level of .com bubble or the subprime mortgage crisis...🤔
@AIBear @ScottAlexander I am being serious, would you count AI bubble burst as "effect of AI" or something else?
@AIBear Thanks for this funny edge case. I think that it is most within the spirit of the question to say that no, that would not be a direct effect of AI (even though it would be an effect of the AI industry).
Going with NO because I see AI being used just enough to maintain trendline growth/productivity. As for unemployment, it has nowhere to go but up, and new technology is not going to be a popular explanation for a higher unemployment regime.
AI might somehow be scapegoated for above-trend inflation, but that wasn't in the question.
@Cactus I guess if the portion attributable to AI would itself be visible if it had happened alone, it counts; otherwise, no.