https://www.nbcnews.com/politics/elections/ai-wedge-issue-midterms-2028-bedfellows-rcna244811

why were so many people voting no on this back in 2023?

@Dulaman I suspect it was based on bad guesses about the technical situation, not the political situation. In 2023 a surprising number of people seemed to have an implicit assumption that short-term technical progress wouldn't be a big deal, talking about near-term AI impacts in terms of 2023-AI. Today the situation is even worse: Some nontrivial fraction of people haven't even kept up with current progress, so their expectations for social impacts are essentially based on 2024-AI.

AI importance seems to have surpassed abortion in this chart:

https://x.com/daniel_271828/status/1992007656555590034

If an AI bubble bursts and causes a financial crisis in 2028, will that count?

I have sold out of my NO position. We will likely see more of this going forward https://www.nytimes.com/2025/11/25/us/politics/ai-super-pac-anthropic.html

If neither major party candidate were explicitly anti- or pro-AI but had opposing views on abortion, would that mean this resolves as NO?

@Xizted Personally, I take it to mean that some major factor going into voting decisions is about the candidates' positions relevant for AI (not just the whole party, and not just pro or con)

I think it's plausible that the current landscape would be sufficient to resolve this YES if there were an election today. Think about how much energy+water usage is discussed, children's use of AI, the copyright issues, the economic implications, job loss, etc etc. I think we'll probably see in the midterms but I'd suspect AI-related interest group funding to exceed abortion interest group funding.

@bens I would estimate that about 10x the number of NYT articles are about AI as about abortion today, fwiw

@bens I would estimate that about 10x the number of NYT articles are about AI as about abortion today, fwiw

@bens That’s an indication that it’s a bigger part of the national conversation, but not necessarily that it’s a bigger political issue

@JimHays That's why I worry about how this will be resolved. It seems unclear, because there are different ways of quantifying it. AI is certainly more novel and interesting, so there's more talk about it, but abortion still has far more single issue voters, and for any given politician, switching their position on abortion would probably be a much bigger deal for them politically.

There's probably more NYT articles about bitcoin than abortion. That doesn't mean bitcoin is a bigger political issue, because the bitcoin discussion is not especially political. Likewise, AI may be more discussed than abortion right now, but most of the discussion is cultural/economic rather than political. Most people probably have a strong opinion on how the government should handle abortion, but few have even thought about how the government should handle AI.

@bens There are coalitions of people in American politics that vote purely on the basis of candidate's positions on abortion. AI is certainly topical, but isn't remotely as politically consequential today.

Self driving cars seem like an increasingly likely vector for this to be true.

By 2028 I expect the current hype will have died down a lot. AI will not have cured cancer, or become sentient, or whatever the fuck Manifold and the SV-adjacent "intelligentsia" fantasizes about. It will be a normal tool, good for some stuff, bad for most other stuff. I use it every day to help me code, but farmers use machines to help them farm and I don't think combine harvesters were ever a hot political topic.

@pietrokc People were killing each other over mechanized textile production in the 1810s until the British government sent in 12,000 troops to arrest the protestors.

John Deere is still pretty politically salient today, it's the most commonly cited reason for the right to repair bills that have been passed in a few states recently.

Neither of those really reached 'as significant as abortion' levels, but if AI can reduce demand for knowledge workers in the US by 10%, that's 10 million people who have lost their livelihoods. I think it's quite plausible that that could push the issue to the top priority

@Frogswap That's interesting (though I have not fact checked it), but as you say people protest things all the time and many small laws exist for many diverse reasons. But abortion was at one time (haven't checked lately) the single issue for O(10%) of voters.

An important requirement for something to be a big political issue is that people have to disagree about it. If 10M people lose their livelihoods, do they disagree on anything? I guess somehow the impacted people could be overwhelmingly Dems or GOP, but afaik "knowledge work" is done ~equally by supporters of both parties.

@pietrokc If 10 million voters think their only chance of holding on to their living conditions is through government action, both parties are going to be advertising their strategy as much as they can going into the 2028 election. I expect that's true even if they have the same position, but there are plenty of degrees of freedom beyond everyone being anti-AI. One might support a ban while the other supports UBI via nationalization, for example.

@pietrokc If the hype dies down a lot, there is quite a chance that would cause a huge market crash (“35% of the US stock market is held up by five or six companies buying GPUs”). Would that count as “AI being a political issue”?

@PetrKadlec in my understanding the crash itself would not count, but it would likely lead to demands for regulation etc which would count

@PetrKadlec I read that essay and I thought it was dumb. I don't think the hype dying down would lead to a huge market crash.