I work at the Alignment Research Center (ARC). I write a blog on stuff I’m interested in (such as math, philosophy, puzzles, statistics, and elections): https://ericneyman.wordpress.com/
Eric Neyman
Karma: 6,317
I probably agree that if you told this sentence to EAs, then they’d be confused or think that things have gone wrong. But like, it’s usually the case that you can give partial information about a situation that leads people to be confused?
I think that if you gave the additional information of “Oh by the way, advanced AI seems imminent, and we are absolutely not on track to handle the situation well, but some candidates with great ideas and a good understanding of AI policy are running for office and could move the needle”, then they would no longer be confused.
I’m not familiar with New York politics in particular, but—at least on the federal level—co-sponsorship doesn’t tell you very much beyond “this person supports the bill”. The RAISE Act had 34 co-sponsors (Lasher among them). Alex, on the other hand, was the driving force behind the bill, helped write the bill, and expended a really large amount of political capital getting it through.
I wouldn’t say that the bridge-burning effect is zero, but I think that raising Alex’s a priori win probability from 5% to 30% was dramatically larger (like, by more than a factor of 100) than the bridge-burning effect, given that Lasher likely won’t be one of the leading representatives on the issue.
There are many people in Congress who favor a data center moratorium. However, none of them have AI as their #1 issue, and this is also true of Lasher. I expect Lasher to be a good member of Congress, but I don’t think Congress will be meaningfully different in its orientation to AI issues because he’s there.
Meanwhile, I think Congress would have been meaningfully different on AI issues if Bores had won. Bores has made AI his #1 issue, and I think he would have done all he could to pursue policies that try to mitigate AI x-risk. He would have staked a lot of political capital on that, as he did with the RAISE Act.
I expect to write a longer reflection, but my position is that this was a really good bet for us to have made, even if it didn’t pan out (and indeed, in my original post, I assigned only a 20% chance to it panning out).
(I also think that the positive effect on the narrative around LTF if Bores had won would have been a very small fraction of the direct positive effect of having Bores in the US government.)
I don’t, sorry! Although, for what it’s worth, in the two conversations I’ve had with him about risks from AI, China didn’t come up.
Edited to add: also, for what it’s worth (and apologies for venturing into explicitly partisan politics), I am personally substantially more worried about AI-enabled authoritarianism under Trump than under most other possible presidents (though my worry definitely doesn’t go away entirely).
I don’t speak for Bores, but I think Eli is correct.
Thanks for posting this, and it’s something I’ve been thinking about, too. I think that running for office is really rough, and—pretty unfortunately—there can be a trade-off between sticking to your guns on your beliefs and winning elections. I’m not sure what to do about this and how I would advise politicians to navigate it.
I think I straightforwardly disagree with this part:
Alex Bores seems like a very mildly below-average integrity politician
I’m not sure exactly where my view (that Alex is an above-average integrity politician) comes from, but here are some pieces of evidence:
During his time in the Assembly, he has sometimes been one of the only Democrats voting against tech regulation bills that didn’t make sense. This is despite those votes being politically inexpedient for him. See here for some examples; my understanding is that A4550 is an example of straightforwadly bad regulation.
I believe that the amount of political capital he staked on fighting for a strong version of the RAISE Act in December is much larger than the most politically opportune amount of capital for him to have staked.
More generally, people who have worked closely with him say that he cares a ton about the substance of bills he fights for, rather than accepting more superficial wins that are politically expedient.
I also want to push back on this part--
If, based on the dozens of people in the AI Safety community writing glorious exhortations of him, you ended up with the impression that when Alex says something you can take it remotely as seriously as posts around here, please update.
--to which I reacted with “agree denotationally but object connotationally”. Here’s what I believe, and I’m curious if you disagree with: I think we have reasonably strong evidence that Bores’s primary concern on AI issues is catastrophic and existential risks.[1]
I have two main pieces of evidence for this:
The people who have worked most closely with him on bills—who, by the way, have dealt with many politicians—believe it to be true.
His defense of the RAISE Act on the State House floor stressed catastrophic risk in a way that actually pushed back on concerns about more immediate harms.
This doesn’t address all of your critiques of Bores, and I’m pretty sympathetic to your critique of his social media presence.
- ^
My guess is that his top two concerns are bio misuse risk and AI-enabled authoritarianism; I believe yours is AI takeover, so I’m not saying that you and Bores are precisely aligned.
A heuristic I’ve heard used is that spending on advertising in primary elections in inexpensive media markets has an effect of roughly $100/net vote. On the other hand, Berkeley is an expensive area to run ads, so it should be higher.
(And this should be treated as a primary election, because it’s a Democrat vs. Democrat contest.)
Sometimes we see races where one side starts spending large amounts all of a sudden and the polls shift; this lets us get an okay-but-not-great measurement of the cost-effectiveness of ads. See eg here (search “Iowa”) for an example—though admittedly that was spending by a super PAC, which is generally somewhat less effective than campaign spending, so you should divide those cost estimates, maybe by a factor of 1.5 or so.
My guess would be something like: starts at $200/vote and then gradually falls off as diminishing marginal returns kick in.
But in terms of my overall appraisal of the situation, I’m operating under a different heuristic: incumbents are very difficult to beat unless they are unpopular. And, as far as I know, Lateefah Simon is perfectly popular.
Jamie raised $14K so far and spent 10k on the primary, winning 28K votes (!!), meaning getting a vote for 36cts, which seems fairly efficient.
For figuring out spending effectiveness, the question is how many votes she would have gotten if she hadn’t spend any money on the primary. My guess is, perhaps, a couple hundred fewer votes.
Thanks for writing this, Jeff!
If anyone is interested in receiving political donation recommendations, feel free to fill out this form!
Do we know that he didn’t just change his mind?
I think voters might not trust that the politicians will in fact vote in accordance with their constituents’ beliefs. It’s hard to credibly commit to that.
Campaign update: With 56% of the vote in, AI safety champion Scott Wiener and Connie Chan (both Democrats) have advanced to the November election. Wiener got 41.3% of the vote, compared to Chan’s 28.6%. This is roughly what I expected.
Wiener is favored to win in November, but it’s not a done deal, for a few reasons:
Voters for the third-place finisher, Saikat Chakrabarti, might go disproportionately to Chan over Wiener: I think Chakrabarti is more ideologically similar to Chan overall, though it’s not totally clear.
Republicans really don’t like Wiener, and might prefer Chan over Wiener.
Nancy Pelosi, who endorsed Chan, will probably use what power she has to try to further boost Chan.
Wiener is about 70% to win per prediction markets, and that sounds about right to me. (Maybe it’s a tad low; my intuition is more like 75%.)
Getting elected involves compromising your values.
Usually, when people say “compromising values”, it carries a connotation of low integrity. That’s not really my intention here. Instead, I mean it in a more neutral way: if you’re running for office, it’s pretty likely that your values will be out of step with your constituents’ values in one way or another. Maybe you have a wider moral circle of concern, or have a stronger sense of justice, or whatever, leading to you holding views that are unpopular among your constituents.
This makes you less likely to win. And so, you have three choices:
Stick to your values, tanking the decreased likelihood of winning
Change your values, or at least commit to acting on your constituents’ values rather than your own
Deceptive alignment: Run on your constituents’ values, but once you’re in power, act on your own values
Options 1 and 2 trade off against each other, and different politicians have reputations for being at different points on the spectrum. For example, Bernie Sanders is thought of as pretty close to 1, while Gavin Newsom is thought of as pretty close to 2.
The most clean-cut example of #3 that I know of is Jimmy Carter, who deliberately tried to appeal to segregationists in his campaign for governor, but then famously declared that “the time for racial discrimination is over” in his inaugural address. I find this interesting to think about, because I think that racial discrimination is bad, but at the same time I think it’s a bad policy to be deceptively aligned to your constituents.
Anyway, I sometimes think about this in the context of electing AI safety champions to Congress. I think the value proposition is great, but the more they take route 2 over route 1, the lower the value of electing them. (Meanwhile, I think that AI safety champions should steer away from route 3, though in my opinion it’s fine to place a different emphasis when campaigning than when policymaking.)
Previously, I’d written about my support for Alex Bores and Scott Wiener, who are running for Congress. I wanted to highlight another person who’s running for office, namely Will Dreher, who’s running for the Washington State House. As far as I know (I haven’t checked thoroughly), he holds the distinction of being the first serious candidate for elected office to put AI as the top issue on his platform.
Dreher’s platform talks about a number of risks from AI, ranging from risks that are already widely discussed in politics (such as deepfakes and job displacement) to catastrophic and existential risks. Based on my conversations with him, it seems to me like he’s taking catastrophic and existential risks from AI pretty seriously and is interested in pushing for legislation that mitigates those risks.
Some quotes from his platform that I like:
“AI companies’ billionaire (soon to be trillionaire) CEOs, meanwhile, alone are deciding pressing public questions like what guardrails will apply to AI’s military applications or whether to release potentially dangerous new models that could pose catastrophic risks, like cyberattacks, loss of control, or bioengineered pandemics. Yet these companies have shown no willingness to slow down—indeed, they have pledged to fight state attempts to regulate them.”
“I will… pass legislation addressing the catastrophic risks that certain AI applications pose, including right-to-warn protections for AI employees who perceive risks from their companies’ work, regulatory audits and controls that apply before even internal deployment of new models, and required transparency by AI companies of risks flagged internally during safety testing.”
I think that Washington is one of the more important states for AI regulation, as a number of AI companies have big offices there. So I’m excited to have an AI safety champion in the Washington legislature. A number of donors who are concerned about catastrophic/existential risk and want to see greater regulation of frontier AI companies have donated to him, and I think further donations are worthwhile. You can donate up to $2.4k here if you’d like (donate via this link, rather than directly through his website).
Mechanistic estimation for expectations of random products
Thanks! I meant to make the narrower point that my probability that the race will be decided by a small number of votes has gone up. I’ve expanded on footnote 8 to clarify how I’ve updated / what has changed.
Yeah. I guess another piece of it is that it’s straightforwardly unpleasant to have attack ads being run against you, regardless of how much they affect your chances of winning.
Note that (I think?) most conjectures are true, though! So you shouldn’t expect 50⁄50 by default.