The results in neighboring races and districts are more troubling—NY elected several lunaticthird worldist socialists to local, state and federal office last night.[1]
I think the main takeaway for EAs and rationalists interested in getting deeper into politics should be to move away from political progressivism as fast as possible. For strategic reasons: the NY results show that progressivism without the destructive third worldism attached is an electoral dead end, and principle / correctness reasons: even “good” progressivism is deeply flawed and unappealing as a political system for addressing AI risk and sociopolitical problems more broadly.
Overall, we find that approximately 55% of alignment researchers identify as politically progressive to some extent, while approximately 80% of EAs identify in the same way.
A lot of AI x-risk-pilled people are wary of polarizing the issue along partisan political lines. But I think (a) it already is polarized somewhat and inevitably will get more polarized, and (b) there’s a good case for polarizing to the right, because the policy prescriptions of (traditional / non-populist) right-wing politics are much less threatening to neutrals / skeptics if AI doom turns out to be false, and also broadly correct on a bunch of other important questions.
IDK how exactly this would fit into a practical / short-term political strategy that takes current trends in US politics into account. It seems likely there will be a blue wave in 2026 / 2028, so perhaps the way to capitalize on that is to support and recruit x-risk pilled moderate democratic candidates to run against incumbent republicans (as opposed to focus on races like NY-12 and CA-11 where it is a foregone conclusion that they will be represented by some kind of democrat). A big challenge with that is geographic though; suitable states and districts to contend in are mostly not where rationalists and EAs live, and it is difficult to recruit carpetbaggers and hard for them to win (and for 2026 specifically it’s too late).
Bernie-backed lefties lost house primaries in NY-12 (Bores) and NY-17, but won in NY-7, NY-10, and NY-13. That’s a respectable win rate and seems promising if you want government action against AI companies. It’s a shame Alex Bores didn’t make it, but he should have an easy enough time getting his newly-elected allies to support his bill.
AFAICT the hard part is getting the larger body of non-left politicians to sign onto harsh AI restrictions when most of them take money from those companies. I suspect the softest target is datacenter construction, because datacenters have been bad neighbors to a lot of people of all political persuasions.
Having talked to voters in NY-12 these past few days, and looking at the cross-tabs of the results, Bores outperformed Lasher in one particular kind of precinct: those that were markedly more progressive. In other words, the close result leading to headlines like this were possible because of Bores’ alignment with progressives. Other analyses indicated that a Bores win would come from turnout among younger and more progressive groups, and in reality, that result just never materialized and turnout among Lasher’s base was stronger.
Additionally, I think the success of the DSA is better seen as a categorical rejection of the status quo by voters rather than categorical acceptance of “third-worldism.” Mamdani ran his campaign for mayor last year on such a rejection, and voters perceive him as delivering on the effective, impactful, and material change EA folk dream about.
You are right that this race was not entirely a referendum on AI x-risk, but it was a referendum on what role AI will play in politics. I think the question at stake was: is meaningful and smart AI regulation something that will be defeated by anti-democratic, monied interests? To me, the close result answers in the negative.
There is a difference between allying with unions, preferring a larger welfare state, generally favoring more heavy-handed govt regulation, etc. (normal progressivism) vs. support for wholesale confiscation of private property, Hamas, and prison / border abolition. In multiple races adjacent to NY-12, candidates who support the latter beat candidates who support the former. If that’s what turns out “young progressives”, no one should be allying with them or trying to appeal to them.
Putting aside whether such descriptions are accurate or whether they could be described as “third-worldism,” this still doesn’t change what I said previously. I doubt the success of these candidates stems purely from their support for these ideas, but rather a rejection of the political establishment.
These same areas also swung towards Trump in 2024 before voting for Mamdani. Sometimes the simplest explanation is the right one!
The Bores loss is disappointing, but NY-12 was a close race between two relatively sane and high-quality candidates, and in actuality not really a referendum on AI x-risk or salience, despite the spending / attention. (more commentary from some rat-adjacent people: https://x.com/peterwildeford/status/2069781365084574098, https://x.com/daniel_271828/status/2069625692271398917, https://x.com/peterwildeford/status/2069821339112763886)
The results in neighboring races and districts are more troubling—NY elected several lunatic third worldist socialists to local, state and federal office last night.[1]
I think the main takeaway for EAs and rationalists interested in getting deeper into politics should be to move away from political progressivism as fast as possible. For strategic reasons: the NY results show that progressivism without the destructive third worldism attached is an electoral dead end, and principle / correctness reasons: even “good” progressivism is deeply flawed and unappealing as a political system for addressing AI risk and sociopolitical problems more broadly.
From a 2024 survey:
A lot of AI x-risk-pilled people are wary of polarizing the issue along partisan political lines. But I think (a) it already is polarized somewhat and inevitably will get more polarized, and (b) there’s a good case for polarizing to the right, because the policy prescriptions of (traditional / non-populist) right-wing politics are much less threatening to neutrals / skeptics if AI doom turns out to be false, and also broadly correct on a bunch of other important questions.
IDK how exactly this would fit into a practical / short-term political strategy that takes current trends in US politics into account. It seems likely there will be a blue wave in 2026 / 2028, so perhaps the way to capitalize on that is to support and recruit x-risk pilled moderate democratic candidates to run against incumbent republicans (as opposed to focus on races like NY-12 and CA-11 where it is a foregone conclusion that they will be represented by some kind of democrat). A big challenge with that is geographic though; suitable states and districts to contend in are mostly not where rationalists and EAs live, and it is difficult to recruit carpetbaggers and hard for them to win (and for 2026 specifically it’s too late).
Technically they won primaries, but in districts where the general election is typically not competitive.
Bernie-backed lefties lost house primaries in NY-12 (Bores) and NY-17, but won in NY-7, NY-10, and NY-13. That’s a respectable win rate and seems promising if you want government action against AI companies. It’s a shame Alex Bores didn’t make it, but he should have an easy enough time getting his newly-elected allies to support his bill.
AFAICT the hard part is getting the larger body of non-left politicians to sign onto harsh AI restrictions when most of them take money from those companies. I suspect the softest target is datacenter construction, because datacenters have been bad neighbors to a lot of people of all political persuasions.
Having talked to voters in NY-12 these past few days, and looking at the cross-tabs of the results, Bores outperformed Lasher in one particular kind of precinct: those that were markedly more progressive. In other words, the close result leading to headlines like this were possible because of Bores’ alignment with progressives. Other analyses indicated that a Bores win would come from turnout among younger and more progressive groups, and in reality, that result just never materialized and turnout among Lasher’s base was stronger.
Additionally, I think the success of the DSA is better seen as a categorical rejection of the status quo by voters rather than categorical acceptance of “third-worldism.” Mamdani ran his campaign for mayor last year on such a rejection, and voters perceive him as delivering on the effective, impactful, and material change EA folk dream about.
You are right that this race was not entirely a referendum on AI x-risk, but it was a referendum on what role AI will play in politics. I think the question at stake was: is meaningful and smart AI regulation something that will be defeated by anti-democratic, monied interests? To me, the close result answers in the negative.
There is a difference between allying with unions, preferring a larger welfare state, generally favoring more heavy-handed govt regulation, etc. (normal progressivism) vs. support for wholesale confiscation of private property, Hamas, and prison / border abolition. In multiple races adjacent to NY-12, candidates who support the latter beat candidates who support the former. If that’s what turns out “young progressives”, no one should be allying with them or trying to appeal to them.
Putting aside whether such descriptions are accurate or whether they could be described as “third-worldism,” this still doesn’t change what I said previously. I doubt the success of these candidates stems purely from their support for these ideas, but rather a rejection of the political establishment.
These same areas also swung towards Trump in 2024 before voting for Mamdani. Sometimes the simplest explanation is the right one!
https://www.thefp.com/p/darializa-avila-chevalier-congress-third-worldism
https://www.thefp.com/p/what-the-right-gets-wrong-about-zohran