Looking back, it seems quite possible to me that EAs undervalued the importance of helping with the previous election. For one limited thing, having this administration be in power when we're getting so close to TAI seems like a major failure.
Personally, I think I've become more convinced recently that generic US policy, especially focusing on long-term issues (like US governance, or US decisions on questions like Nuclear/bio/AI) might be a good use of EA funds.
I would have really guessed that a lot of this area wouldn't at all be neglected, but in practice, it seems to be far more ignored than I think is reasonable.
I agree that US policy is obviously very important, but
The 2024 US Election doesn't seem to have been close (unlike e.g. the 2000 one). I'm skeptical that EAs could have changed things that much, given that they couldn't cause relatively easier policy changes (e.g. SB 1047)
They were a minority, but several EAs favoured Trump over Harris, so their efforts would partially cancel out
Politicizing EAs, or EA organizations, would have made them less effective after Trump's win and the whole "vibe shift", when collaborating with the current administration
generic US policy, especially focusing on long-term issues (like US governance, or US decisions on questions like Nuclear/bio/AI) might be a good use of EA funds.
I think SBF was fairly unique in caring about this, and his empire collapsed before he did anything like this in that election. When I said "undervalue", I wasn't referring to SBF, given he hasn't been active in this time. (Obvious flag that while I might have sympathized with some of SBF's positions, I very much disagree with many of his illegal and fraudulent actions)
Looking back though, paying $5B for Trump to definitely not run/win seems like a great deal to me, though the act of paying him raises a lot of qualms I'd be uncomfortable with.
And we should also examine neglect not just on the headline number of dollars going into the space but on specific facets, like how much money in that space actually goes to top tier impact opportunities or how much investment is there in innovating the space/interventions.
I think in-comparison to the space, EA has a comparative advantage more in talent than in money. I think the Harris campaign got $2B or so of donations, but I get the impression that it could have used smarter + more empirically-minded people. That said, there is of course the challenge or actually getting those people to be listened to.
I have not seen a lot of evidence that EA skills are very transferable to the realm of politics. As counterexamples, look at the botched Altman ouster, or the fact that AI safety people ended up helping start an AI arms race: these partially seem to come from a place of poor political instincts. EA is also disproportionately STEM background, which are generally considered comparatively poor at people skills (accurately, in my experience).
I think combating authoritarianism is important, but EA would probably be better off identifying other people who are good at politics and sending support their way.
My impression from other writing I've seen of yours is that you don't think that EAs are good at too many things. What do you think EAs are best at, and/or should be doing? Perhaps, narrow GiveWell-style research on domains with lots of data?
I think you probably need multiple kinds of skill and some level of cognitive style diversity within a political campaign. You definitely need a lot of people with people skills, and I am sure that the first gut instincts of people with good social skills about what messaging will work are better than those of people with worse social skills. Those socially skilled people should undoubtedly be doing detailed messaging and networking for the campaign. But you also need people who are prepared to tell campaigns things they don't want to hear, even when there is severe social pressure not to, including things about what data (rather than gut instinct) actually shows about public opinion and messaging. (Yes, it is possible to overrate such data which will no doubt be misleading in various ways, but it also possible to underrate it.) My guess is that "prepared to tell people really hard truths" is at least somewhat anticorrelated with people skills and somewhat correlated with STEM background. (There is of course a trade-off where the people most prepared to tell hard truths are probably less good at selling those truths than more socially agreeable people.) For what it's worth Matt Yglesias' seems pretty similar to the median EA in personality, and I recall reading that Biden advisors did read his blog. Ezra Klein also seems like a genuinely politically influential figure who is fairly EA-ish. There is more than one way to contribute to a political movement.
I personally don't think EA should be doing much to combat authoritarianism (other than ideally stopping its occasional minor contributions to it via the right-wing of rationalism and being clear-eyed about what the 2nd Trump admin might mean for things like "we want democratic countries to beat China") because I don't think it is particularly tractable or neglected. But I don't think it is a skill issue, unless you're talking about completely EA run projects (and even then, you don't necessarily have to put the median EA in charge; presumably some EAs have above average social skills.)
(other than ideally stopping its occasional minor contributions to it via the right-wing of rationalism and being clear-eyed about what the 2nd Trump admin might mean for things like "we want democratic countries to beat China")
Actually I think this is the one thing that EAs could realistically do as their comparative advantage, considering who they are socially and ideologically adjacent to, if they are afraid of AGI being reached under an illiberal, anti-secular, and anti-cosmopolitan administration: to be blunt, press Karnofsky and Amodei to shut up about "entente" and "realism" and cut ties with Thiel-aligned national security state companies like Palantir.
I don't think cutting ties with Palantir would move the date of AGI much, and I doubt it is the key point of leverage for whether the US becomes a soft dictatorship under Trump. As for the other stuff, people could certainly try, but I think it is probably unlikely to succeed, since it basically requires getting the people who run Anthropic to act against the very clear interests of Anthropic and the people who run it (And I doubt Amodei in particular, sees himself as accountable to the EA community in any way whatsoever.)
For what it's worth I also think this complicated territory and that there is genuinely a risk of very bad outcomes from China winning an AI race too, and that the US might recover relatively quickly from its current disaster. I expect the US to remain somewhat less dictatorial than China even in the worst outcomes, though it is also true that even the democratic US has generally been a lot more keen to intervene, often but not always to bad effect, in other country's business.
Conditional on AGI happening under this administration, how much AGI companies have embedded with the national security state is a crux for the future of the lightcone, and I don't expect institutional inertia (the reasons why one would expect "the US might recover relatively quickly from its current disaster" and "the US to remain somewhat less dictatorial than China even in the worst outcomes") to hold if AGI dictatorship is a possibility for the powers that be to reach for.
It increases the AI arms race thus shortening AGI timelines, and, after AGI, increases chances of the singleton being either unaligned or technically aligned to being an AGI dictatorship or other kind of dystopian outcome.
My knee-jerk reaction is to feel attacked by this comment, on behalf of the EA community.
I assume that one thing that might be going on is a miscommunication. Perhaps you believe that I was assuming that EAs could quickly swoop in, spent a little time on things, and be far more correct than many experience political experts and analysts.
I'm not sure if this helps, but the above really doesn't align with what I'm thinking. More something like, "We could provide more sustained help through a variety of methods. People can be useful for many things, like direct volunteering, working in think tanks, being candidates, helping prioritization, etc. I don't expect miracle results - I instead expect roughly the results of adding some pretty smart and hardworking people."
I instead expect roughly the results of adding some pretty smart and hardworking people.
The usefulness of smart people is highly dependent on the willingness of the powers-that-be to listen to them. I don't think lack of raw intelligence had much of anything to do with the recent US electoral results. The initial candidate at the top of the ticket was not fit for a second term, and was forced out too late for a viable replacement to emerge. Instead, we got someone who had never polled well. I also don't think intelligence was the limiting factor in the Democrats' refusal to move toward the center on issues that were costing them votes in the swing states. Intellectually understanding that it is necessary to throw some of your most loyal supporters under the bus is one thing; committing to do it is something else; and actually getting it done is harder still. One could think of intelligence as a rate-limiting catalyst up to a certain point, but dumping even more catalyst in after that point doesn't speed the reaction much.
I think @titotal's critique largely holds if one models EAs as a group as exceptional in intelligence but roughly at population baseline for more critical and/or rate-limiting elements for political success (e.g., charisma, people savvy). I don't think that would be an attack -- most people are in fact broadly average, and average people would be expected to fail against Altman, etc. And if intelligence were mostly neutralized by the powers-that-be not listening to it, having a few hundred FTEs (i.e., ~10% of all EA FTEs?) with a roughly normal distribution of key attributes is relatively unlikely to be impactful.
Finally, I think this is a place where EA's tendencies toward being a monoculture hurts -- for example, I think a movement that is very disproportionately educationally-privileged, white, STEM focused, and socially liberal will have a hard time understanding why (e.g.) so many Latino voters [most of whom share few of those characteristics] were going for Trump this cycle and how to stop that.
On EAs in policy, I'd flag that: - There's a good number of people currently working in AI governance, Bio governance, and animal law. - Very arguably, said people have had a decent list of accomplishments and power positions, given that such work was fairly recent. See Biden's executive orders on AI, or the UK AI Security Institute. https://www.aisi.gov.uk/ - People like Dustin Moskovitz and SBF were some highly prominent donors to the Democratic party.
I think the EA policy side might not get a huge amount of popularity here, but it seems decently reputable to me. Mistakes have been made, but I think a decent report on the wins and losses would include several wins.
I do agree that finding others doing well and helping them is one important way to help. I'd suspect that the most obvious EA work would look like prioritization for policy efforts. This has been done before, and there's a great deal more that could be done here.
In fairness, SBF was also secretly a prominent Republican donor, right? Didn't he basically suggest in the infamous interview with Kelsey Piper that he was essentially cynical about politics and just trying to gain influence with both parties to help advance FTX and Alameda's interests?
He was a Republican donor, but from what I understand, not really a MAGA donor. My impression was that he was funding people on both sides, who were generally in favor of their interests - but that their interests did genuinely include issues like bio/ai safety.
I think it's very reasonable to try to be bipartisan on these issues.
Fair point. I certainly don't think it is established (or even more than 50% likely) that SBF was purely motivated by narrow personal gain to the exclusion of any real utilitarian convictions at all. But I do think he misrepresented his political convictions.
It is orthogonal. More that TAI might be soon, we probably want an administration that would both promote AI safety and broadly be cooperative/humble/deliberate.
I have argued previously that A) liberal democracy and stability of the American system is under threat and B) the trajectory of American political dysfunction and polarization is unsustainable and we’ve passed theoretical red lines where a course correction would have happened if there was going to be one.
I don’t have an updated version of this piece for the 2025, but I’ll link to this briefing that thoroughly catalogs authoritarian probing and state cannibalization.
Here is the high-level case for tractability: The EA+EA-adjacent sphere has improved and arguably disrupted the (philanthropic) global health/development and animal welfare spaces yielding major impact. When looking at democracy/“resistance” space as a whole, I think there is a clear case that EA could fill a similar role by A) pushing the effectiveness mindset/consequentialist thinking, B) building and promoting GiveWell-like orgs in the space (see Power for Democracies and Focus for Democracy), C) incubating envelope-pushing interventions similar to Charity Entrepreneurship (Movement Labs might be the closest reference in the democracy space), D) bringing in more funding and funding that is also more comfortable in hits-based giving (Democracy Fund’s report reflects the ability to absorb more funding and innovation). I think an evolution in the democracy space similar to the ongoing evolution in GHD and AW would significantly improve the space’s position to avert authoritarian consolidation and other antidemocratic outcomes.
I will note that GHD and AW are not 1:1 correlations with democracy fragility as cause areas. Unlike GHD and AW, you are staring down a game-over scenario where democracy fails in an irreversible way which makes democracy fragility a more time-sensitive field.
Democracy work is also heavy on the complex systems interactions that are hard to quantify, making it unwise to rely only on a small set of cost-effectiveness recommendations like GiveWell does. In that regard, an org like Democracy Funders Network can be a complement to a GiveWell-style Focus for Democracy.
I am currently helping to develop the posture towards the democracy space of an EA-aligned philanthropic advisory group. When appropriate in the future, I will provide some comments/recommendations regarding the brainstorm you are working on. But I’ll leave readers with my framework and analogy that I use to conceptualize the problem (this is my original analogy, so I’d appreciate attribution if anyone reuses it):
Spectrum of causation:
Upstream e.g. money in politics, first-past-the-post X partisan primaries X single-member districts
Midstream e.g. capture of a major party by an aspiring authoritarian demagogue
Downstream e.g. authoritarian candidates win elections, limited repercussions for insurrection
Immediate threats of authoritarian consolidation e.g. WH administration creating a constitutional crisis by ignoring court orders and testing resolve of judicial branch
Analogy (still a work in progress)
Gas = upstream causes
Flammable material sitting around/fire breaks = midstream
Inaccessible portals/access points = downstream
Flames on load-bearing walls = Immediate threats of authoritarian consolidation
The house, America, is on fire! A gas leak ignited into a raging blaze. The leak went undetected for a long time, though more and more occupants were noticing the strange smell just before the fire erupted, with some even trying to address the apparent leak. Now, the fire threatens to consume the entire house.
The fire hasn’t spread to the entire house but it appears like it could quickly. At the moment it’s threatening some critical load-bearing walls and some portals necessary for firefighters to access certain rooms. Amidst all this, the gas is still leaking and continuing to fuel the fire.
Right now people are frantically trying to douse water on the fire; some are indiscriminately throwing water on the flames closest to them, others are using their water to regain strategic entry points, and the water of some is being used preserve the load-bearing walls to prevent the structure from collapsing, which would render all efforts null. It’s unclear if the load-bearing walls will collapse in, how quickly they could, and which ones are most liable to do so.
Some people are in the house trying to cut firebreaks—removing flammable materials and closing doors—to slow the spread before it reaches untouched rooms; yet there is a lack of clarity on how effective the efforts have been and what rooms to prioritize.
Nobody has turned off the gas in the basement yet, and the fire won’t be truly extinguished until the gas leak is stopped. Efforts to reach the basement in the burning house have had incremental success thus far, and some people are trying to problem solve how to make the treacherous journey to the basement in a house that is on fire. However, this has the least attention at the moment, just as the gas leak did before the fire began.
There is a clear need to triage and be strategic that must be balanced with urgency and the inability to have full confidence in the crisis such as this.
***
My prior is that EA could generally focus more on the bookends of that spectrum.
Tackling immediate threats is pressing because the aggressive onslaught of authoritarian probing appears to be creating a lot of hinge points where either checks and balances work or authoritarian consolidation happens (which this early on in the term is quite bad for free/fair elections and peaceful transfer of power in 2028/2029).
EA would play to its strengths by working to on the upstream causes which are relatively neglected. We got into this situation because society neglected the upstream causes, these causes will continue to be neglected whilst a crisis is perceived, innovation is needed to effectively tackle these upstream causes.
I'm not sold, and I'm going to lay out some general reasons.
Firstly, a lot of the concerns expressed here I think are extremely unlikely. I do not think there is any serious risk that Trump will send the military after, or otherwise seriously harass, former government employees. I do not think he will pursue a third term or otherwise interfere with our tradition of free and fair elections. And I do not think he will openly defy a court order.
Some of the other things you fear I don't necessarily see as bad. As a matter of democratic accountability, by which I mean accountability to the people rather than checks and balances or "good" governance, I do think the president has the right to fire executive branch employees, whether or not we like the particular decisions he makes.
Secondly, and my error bars on this point do cross the zero line, but my expectation is that Trump will reduce the risk of a nuclear or biological catastrophe. The wars in Ukraine and Israel both started on Biden's watch. With Trump I am hopeful that both will reach some kind of resolution on somewhat favorable terms.
I do think it is good that people are filing lawsuits challenging the questionably legal things Trump is doing. I don't think that this intervention is particularly neglected.
Also, you seem to suggest that Danielle Sassoon's actions in regard to the Eric Adams case are somehow an instance of legislative checks on the executive. I don't get that. Sassoon was an employee of the executive branch, not the legislative. That's why an executive branch official was able to fire her.
I haven't thought about this in great depth, so I'm very open to the possibility that this topic should be deprioritised. I haven't understood your rationale, so I hope you don't mind if I probe further.
Firstly, a lot of the concerns expressed here I think are extremely unlikely. I do not think there is any serious risk that Trump will send the military after, or otherwise seriously harass, former government employees.
I guess I'd be somewhat interested to know why serious harassment is so unlikely. The sources that I cited seemed to be quite worrying to me on this front.
The Guardian reported the following: “Trump’s escalating threats to pervert the criminal justice system need to be taken seriously,” said the former justice department inspector general Michael Bromwich. “We have never had a presidential candidate state as one of his central goals mobilizing the levers of justice to punish enemies and reward friends. No one has ever been brazen enough to campaign on an agenda of retribution and retaliation.” And NPR reported that "Trump has issued more than 100 threats to investigate, prosecute, imprison or otherwise punish his perceived opponents".
Having said that, the point I was making relied less on whether Trump would actually seriously harass people, but rather whether they would fear that Trump would do so, and specifically fear this enough that they would avoid taking actions which might act as a check/balance on presidential power. Do you believe that people don't have this fear?
Some of the other things you fear I don't necessarily see as bad. As a matter of democratic accountability, by which I mean accountability to the people rather than checks and balances or "good" governance, I do think the president has the right to fire executive branch employees, whether or not we like the particular decisions he makes.
I'm not sure I follow. Which are things which I fear, but which you don't see as necessarily bad? When I first read this, I thought you were referring to my list of things I fear:
Evisceration of aid becoming permanent
Increased risk of conflict, potentially moving beyond the likes of Greenland and escalating to great power conflict
Increased risk of (accidental or deliberate) use of nuclear weapons. (Apparently the administration fired over 300 employees at the national nuclear security administration, then tried to reinstate them, but at time of writing doesn’t seem to know how; sources: 1,2,3)
Exacerbation of climate change
An unwillingness to follow international norms may lead to greater willingness to develop biological weapons
If tech billionaire “oligarchs” prefer greater deregulation of AI, this could exacerbate the risk of loss of control of AI/misalignment
The human rights abuses typical of a totalitarian state
I'm assuming you do consider all of these to be bad.
When you spoke about the right to fire executive branch employees, were you referring to my concerns about the erosion of democratic institutions? In that section, I observed that:
Trump wants to fire the director of the Office of Government Ethics (OGE) which monitors conflicts of interests. (Source: MSNBC)
Trump fired 17 inspectors general, whose role is to audit the actions of government.
I'm perfectly willing to believe he has that right, but my question is more about whether it leads to better outcomes. Will government make better decisions without the OGE monitoring conflicts of interest? Will government make better decisions if the inspectors general are loyalists? (assuming that's what they are.) I imagine this leads to worse outcomes, but if you are more sanguine I'd be interested to know why.
I do think it is good that people are filing lawsuits challenging the questionably legal things Trump is doing. I don't think that this intervention is particularly neglected.
I had the intuition that there was probably a lot of work that could be done here, but that the firehose of actions meant that it was hard for people to spare the attention on any of them. This gave me the impression that while lawsuits were happening, there's probably lots more that can be done. Not least because lawsuits are often expensive, and could peter out or become ineffective because of lack of funds. This is pretty impressionistic though, so if you have a more carefully researched opinion, I'd be interested.
I guess I'd be somewhat interested to know why serious harassment is so unlikely. The sources that I cited seemed to be quite worrying to me on this front.
The Guardian reported the following: “Trump’s escalating threats to pervert the criminal justice system need to be taken seriously,” said the former justice department inspector general Michael Bromwich. “We have never had a presidential candidate state as one of his central goals mobilizing the levers of justice to punish enemies and reward friends. No one has ever been brazen enough to campaign on an agenda of retribution and retaliation.” And NPR reported that "Trump has issued more than 100 threats to investigate, prosecute, imprison or otherwise punish his perceived opponents".
I think a lot of where we differ is in how much we trust the media when it comes to Trump. I've generally found that media will report ordinary things as though they were extraordinary and bad, will take the worst possible interpretation of ambiguous quotes, and do whatever else they think will keep people irrationally afraid of Trump. Take the particular claim that you made your centerpeice - that Musk was throwing around his wealth to support Trump's nominees. "Rich American uses wealth to influence politicians" is not exactly news - that happens every day on both sides of the aisle. And what you put in block quotes was just that fact wrapped in hyperbolic language. I looked very briefly at your seven sources. All were during the election and all seemed to draw on the same quote: "Trump said that if 'radical left lunatics' disrupt the election, 'it should be very easily handled by — if necessary, by National Guard, or if really necessary, by the military.'" The author somehow read that as "Trump has expressed support for using government force against domestic political rivals." Talk about a straw man! Trump was suggesting using the military to ensure the election happens. That may be unwise and unprecedented and predictably totally unnecessary, but it is nothing like using the military against political rivals. This is the press's standard MO when reporting on Trump. So when they say "100 threats to whatever" I just assume that few if any of the things on their list are actually what they are claimed to be.
Having said that, the point I was making relied less on whether Trump would actually seriously harass people, but rather whether they would fear that Trump would do so, and specifically fear this enough that they would avoid taking actions which might act as a check/balance on presidential power. Do you believe that people don't have this fear?
I agree that people have that fear. I do not think it is warranted. And I think indulging an unwarranted fear is generally a bad idea - you just incentivize people to have unwarranted fears in the future. We need political rhetoric to cool down right now, not heat up.
For your bulleted list of bad things, I agree that many of them are unqualifiedly bad. A few of them I have more nuanced views on. But I don't want to go point by point through it, as I don't think that would shed any light on whether democracy is at risk or what we should do about it.
With regard to IGs and OGE, I'm not too familiar with these institutions, but my read is that the practical implications of who holds these posts are basically nil. IGs don't directly remedy anything, they just write reports. They weren't going to accomplish anything Trump's appointees didn't want them to accomplish anyway.
I get worried when I see people questioning whether the president has the right to fire them, because I value democracy in the literal sense - accountability to the people - and the mechanism by which any executive branch employee is accountable to the people in our system is through the president's ability to fire them.
As for the neglectedness of lawsuits, I think we need to ask about particular lawsuits. It is certainly true that some lawsuits can fail for lack of money or talent. I don't know of a reason to think that either is in short supply when it comes to challenging Trump. As you've pointed out, people are scared as shit, and there are plenty of liberals in the legal profession. But if we want to make the neglectedness case, I want to see it at the level of a particular legal issue, not just "trump is bad for [checks and balances and good government]".
This type of thing is talked about from time to time. The unfortunate thing is that there aren't a ton of plausible interventions. The main tool we have to fight against authoritarianism in the US is lawsuits, and that's already being done and not any place where EA could have a comparative advantage. The other big thing that people come up with is helping Democrats win elections, and there are people working on this, although (fortunately) elections are really ultimately decided by the voters, campaign tactics have limited effect at least at the national level. Besides this I think the most plausible intervention is probably changing election law at the state level though lobbying/advocacy or petitioning for ballot measures - and even there you'd have to find useful measures that are passible (mandating election counting be done on election night so that there's less suspicion of fraud? Giving less leeway to election boards so that they aren't an easy target for theft? score voting?).
PS: voting rights are pretty much a non-issue. The partisan effect of restrictive voting laws is quitesmall, and if if anything these laws probably hurt Republicans these days because they do better among disengaged voters.
The author was transparent about this, with the quote that you highlight. I feel like when the author is clear on the source, and has provided some amount of effort of oversight, then the information could be handled responsibly by readers.
I think including this sort of section is basically a bait and switch - it allows the author to assert positions, potentially persuading people of them, but if challenged on the LLM slop to hide behind "oh yeah but I said I didn't mean that".
I have now reviewed and edited the relevant section.
My feeling when I drafted it was as per Ozzie's comment -- as long as I was transparent, I thought it was OK for readers to judge the quality of the content as they see fit.
Part of my rationale for this being OK was that it was right at the end of a 15-page write-up. Larks wrote that many people will read this post. I hope that's true, but I didn't expect that many people would read the very last bits of the appendix. The fact that someone noticed this at all, let alone almost immediately after this post was published, was an update for me.
Hence my decision to review and edit that section at the end of the document, and remove the disclaimer.
I actually like that you did this; there's such little information on the news "firehose" right now that a possible accuracy/content tradeoff is entirely reasonable!
Seems notable that my modes is that OpenPhil has stopped funding some right wing or even centrist projects, so has less power in this world than it could have done.
Most philanthropic (vs. government or industry) AI safety funding (>50%) comes from one source: Good Ventures, via Open Philanthropy.2 But they’ve recently stopped funding several categories of work (my own categories, not theirs):
Likewise almost all the people I know 2nd hand who have power in the new admin are via either twitter or rationality, both spaces where EA has, to my eye, sought to distance itself in recent years.
If Trump were able to substantially erode constitutional checks and balances, it seems pretty clear to me that the consequences have the potential to be bad, the only question is how bad, and how likely is it.
It is unclear to me whether less democracy would increase or decrease economic growth, which has been very connected to human welfare. So I do not know whether less democracy would increase or decrease human welfare.
To illustrate how bad the outcomes might be, here’s a brainstormed list of potential outcomes which may or may not happen
The outcomes you listed are bad for humans, but I am not confident they are bad for animals, and I guess the effects on animals are the driver of the overall effect.
Exacerbation of climate change
I believe there is significant uncertainty about whether increasing global temperature is good or bad to humans (farmed animals, and wild animals).
"It is unclear to me whether less democracy would increase or decrease economic growth, which has been very connected to human welfare. So I do not know whether less democracy would increase or decrease human welfare."
I usually think your posts are very good because you are prepared to honestly and clearly state unpopular beliefs. But this seems a bit glib: economic growth is not the only thing that effects well-being, by any means, and so simply being unsure about how democracy effects it is not a strong case on its own for being unsure whether democracy increases or decreases human well-being. Growth might be the most important thing of course, but if you really are neutral on the effect of democracy on growth, other factors will still determine whether you should think democracy is net beneficial for humans in expectation.
Also, in the particular case of the US to evaluate whether democracy continuing is a good thing for human well-being, what primarily matters is how democracy shapes up versus the realistic alternatives in the US, not whether democracy is the best possible system in principle, or even the best feasible system in most times and places. It's not like we are comparing democracy in the US to the Chinese communist system, market anarchism, sortition or the implementation of the knowledge-based restrictions on the franchise suggested by Jason Brennan in his book Against Democracy. We are comparing it to "on the surface democracy, but really Musk and Trump use the justice department to make it impossible for credible opponents to run against the Republican party for many national offices or against their favoured candidates in crucial Republican primaries, and also Musk can in practice stop any government payment to anyone so long as Trump himself doesn't prevent him doing so." Maybe you think the risk of that is low, but that's what people are worried about. Maybe you also think that might be good, because Republican policies might be better for growth and that dominates all other factors, but even then, it's worth being clear about what you are advocating agnosticism about and its not the merits of democracy in the abstract, but the current situation in the US.
But this seems a bit glib: economic growth is not the only thing that effects well-being, by any means, and so simply being unsure about how democracy effects it is not a strong case on its own for being unsure whether democracy increases or decreases human well-being. Growth might be the most important thing of course, but if you really are neutral on the effect of democracy on growth, other factors will still determine whether you should think democracy is net beneficial for humans in expectation.
Holding real gross domestic product (real GDP) per capita constant, there is significant cross-country variation respecting whether more democracy and human rights are associated with a smaller/larger population size, life expectancy at birth, share of people who are satisfied with their life, and life satisfaction. Moreover, I would still wonder about causality even if there was a robust positive association between more democracy and human rights and positive social outcomes holding real GDP per capita constant.
I am pessimistic about finding robust causal relationships between democracy and social outcomes. My sense is that the relationship between democracy and growth is among the ones which have been studied the most, and it looks like there is still lots of uncertainty. I think focussing on more fine-grained elements of democracy would be good for this reason. One may have little idea about whether democracy increases or decreases growth in general, but still have a good sense of how concrete policies more often proposed by Democrats or Republicans affect growth.
Also, in the particular case of the US to evaluate whether democracy continuing is a good thing for human well-being, what primarily matters is how democracy shapes up versus the realistic alternatives in the US, not whether democracy is the best possible system in principle, or even the best feasible system in most times and places.
Great point. It is not as if there is a lever one can pull to change the overall level of democracy. I think it is more productive to ask about how concrete policies would change human, animal or non-biological welfare relative to the counterfactual policies.
Maybe you also think that might be good, because Republican policies might be better for growth and that dominates all other factors, but even then, it's worth being clear about what you are advocating agnosticism about and its not the merits of democracy in the abstract, but the current situation in the US.
I am not familiar with the situation in the US, so I am staying close to my libertarian prior. I think my views are close to Bryan Caplan's with respect to the policies that increase human welfare. However, I am much more uncertain/agnostic overall because I care about animal welfare too.
Fair enough, I actually think it is very hard to discover causal relationship in any social scientific domain. I still strongly suspect that dictatorial governments are bad however. (It's almost impossible to get data on the effects of highly developed countries by modern standards ceasing to be democracy, because this has almost never happened.)
Unclear what (economic) libertarianism implies about the Trump admin. They will cut taxes, but also they might put up tarifs.
It does sound sort of interesting, but I don't think I have a clear picture of the theory of change. How does the dashboard lead to better outcomes? If the theory of change depends on certain key people (media? Civil servants? Someone else?) making use of the dashboard, would it make sense to check with those people and see if they would find it useful? Should we check if they're willing to be involved in the creation process to provide the feedback which helps ensure it's worth their while to use it?
I think it would allow many very online slightly anxious people to note how the situation is changing, rather than plug their minds into twitter each day.
I think the bird flu site helped a little in my part of twitter to tell people to chill out a bit and not work themselves into a frenzy earlier than was necessary. At least one powerful person said it was cool.
I think that civil servants might use it but I'm not sure they know what they want here and it's good to have something concrete to show.
Shortly after I wrote this, the news reported nationwide protests on topics pretty aligned you what in talking about here. This might mean that my assessment of neglectedness should be updated
This is a crosspost from the new Animal Welfare Alignment Newsletter by Anima International. You can subscribe on Substack if you are interested in following these efforts. Audio reading also available on Substack.
The goals of this post are to:
1. Raise a question I see as crucially important to the goal of aligning AI to animal welfare...
“How long have you been v*g*n?”
This is one of the most common icebreakers at animal protection events. It’s a baseline assumption, and it mostly holds true: if you’re out advocating for animals not to be tortured or abused, realistically these days you are v**n, or close. And it makes for good conversation. It seems fairly safe to assume when you meet strangers.
But this assumption is hurting the movement in a way which we don’t always notice: someone new comes into the sp...
AI Use Note: Main body text entirely human written. Claude (Opus 4.8) helped develop models of animal life histories in the appendix.
Cross-posted from Good Structures.
Executive Summary
* Animal advocates sometimes make claims like “there are X of this animal...
Status of this post: in writing this I prioritised speed over quality. I'm not particularly expert on this topic. I fully expect there to be things I’m missing or have got wrong. This post is more a request for feedback than a statement of a high-conviction claim.
Summary
I think it's moderately likely that action to bolster American democracy may be high impact.
Importance (“I of ITN”): this seems high to me. For reasons of brevity, I haven’t carefully defended this claim.
Tractability (“T of ITN”): unclear. I’ve set out some thoughts, and it seems there are several things that can be done (either with people’s time or their money). I haven’t taken the time to assess the quality of these ideas, but I have hope that at least some proportion of them are effective.
Neglectedness (“N of ITN”): moderately high, I think. There are people doing things to support the quality of American democracy, but it seems to me that the amount of work happening is low compared to the level of threat. This may be because the firehose of actions has been so full on that people/civil society are too overwhelmed to respond.
My main ask of readers is to express a view on how effective this area of work is. Specifically, I think it would be useful to review the actions outlined in the section “T of ITN”, and form a view on which actions are most effective, and whether they are effective enough to suggest we should act on those (instead of other things that we might currently be donating to or otherwise supporting). If the conclusion is that we should prioritise other things over this problem, I predict it will be because we can’t realistically achieve good traction, especially since I haven't taken the time to assess the effectiveness of the actions that I listed.
When I say “democracy” I roughly mean good governance
Although I’ve used the term “democracy”, what I care about is
Good governance
Checks and balances
I’m using “democracy” here largely because it’s a shorthand which is meaningful for most people. Also, I don’t know of a system which is effective at achieving those goals other than democracy.
Ultimately, these checks and balances protect a country from descending into the worst excesses of authoritarianism.
To what extent is democracy under threat in the US?
The first few weeks of the Trump presidency have seen a firehose of executive orders and other actions. Any one of them on their own would at least raise eyebrows, if not provoke outrage. Several actions appear unconstitutional or at least designed to weaken democracy, but are nonetheless being enacted. So it appears that American democracy is under threat. As an appendix, I’ve put together several such signs from recent events.
I of ITN: Importance seems high
If Trump were able to substantially erode constitutional checks and balances, it seems pretty clear to me that the consequences have the potential to be bad, the only question is how bad, and how likely is it. To illustrate how bad the outcomes might be, here’s a brainstormed list of potential outcomes which may or may not happen:
Evisceration of aid becoming permanent
Increased risk of conflict, potentially moving beyond the likes of Greenland and escalating to great power conflict
Increased risk of (accidental or deliberate) use of nuclear weapons. (Apparently the administration fired over 300 employees at the national nuclear security administration, then tried to reinstate them, but at time of writing doesn’t seem to know how; sources: 1,2,3)
Exacerbation of climate change
An unwillingness to follow international norms may lead to greater willingness to develop biological weapons
If tech billionaire “oligarchs” prefer greater deregulation of AI, this could exacerbate the risk of loss of control of AI/misalignment
The human rights abuses typical of a totalitarian state
For brevity, I have not taken the time to assess the likelihood of these outcomes. Intuitively, it seems that the score for the I of ITN should be at least moderately high, and likely very high. If anyone disagrees with this, I’d be interested to hear why.
T of ITN: Several things can be done, but how effective are they?
In this section, I’ve brainstormed several ideas for interventions. The key question, in my view, is whether the best of these interventions is sufficiently effective.
Here’s my thinking on how to categorise interventions: success occurs when we prevent (or reduce the harm of) illegal/questionable acts conducted by the President, or (more broadly) the executive branch. Let’s structure the options between the following three categories:
Checks and balances: controlling illegal/questionable presidential activity when it happens.
Responsive: actions to perform after illegal/questionable presidential activity, including lawsuits, supporting the media, and citizen action.
Preventative: preventing future illegal/questionable presidential activity, including enabling future elections to be free and fair elections.
This isn’t necessarily a “clean” division (arguably the whole point of some responsive actions is typically that they have a preventative effect).
In each section, I list some organisations which can be supported through donations or volunteering. In each case, I have done almost zero research of these organisations. To a large extent, my call to action is not necessarily to immediately donate/volunteer, but rather to crowdsource views on the effectiveness of taking these actions.
Checks and balances
This section focuses on the checks and balances built into the constitution.
My understanding is that the legislative branch, ie Congress, should play a key role in this, as should the judicial branch.
To a certain extent, this is happening at the moment. See, eg, the recent Eric Adams case (sources 1,2,3, with the actions of Danielle Sassoon being particularly interesting). We see brave individuals (lawyers, in this case) who have demonstrated that they are willing to take brave action to uphold the rule of law.
Although the checks and balances are working at the moment to some extent, it may be that they are working less well than they should. Some sources (eg The Hill, Ezra Klein) argue that Elon Musk is influencing Congress by threatening to use his wealth to unseat members of Congress who defy Trump.
An example which suggests that this threat is working may be the recent controversial appointments, such as Pete Hegseth and RFK. I found this section from an NBC news article on Pete Hegseth instructive:
Joni Ernst, a key Republican vote, plays nice
As a combat veteran and sexual assault survivor, Sen. Joni Ernst, R-Iowa, has long been seen as a linchpin of Hegseth’s path to 50 votes, as she has expressed concerns about him.
At Tuesday's hearing, she did little to challenge Hegseth. And after it ended, Ernst announced her support for his nomination.
<...>
Ernst, who faces re-election in 2026, had come under heavy pressure from Trump allies to get behind Hegseth after her initial skepticisms of his nomination came to light. Lately, she hasn’t shown much of an appetite lately to buck the president-elect.
Furthermore, it seems the ramifications of defying the President could go beyond simply losing their job. For example, there are signs he’s willing to use legal harassment against his opponents (sources: 1,2,3) and even willing to use the US military against his political enemies (sources: 1,2,3,4).
This suggests that work which strengthens the judiciary and legislative branches ability to serve as a check/balance on the executive branch is useful, both at the “losing your job” level of threat and the (probably worse) “legal and military harassment” level of threat.
I expect that there are probably lots of ideas that I’m missing, and would invite more suggestions.
Activity
Brainstormed (ie not at all carefully researched) NGOs/orgs working on this
Defence against threat of being unseated by Musk
Americans may wish to find ways to support members of Congress who may defy the President. This could involve writing in support, donating, or campaigning.
Defence against legal and military harassment
Peace Brigades International (PBI) – A global NGO that provides nonviolent protective accompaniment to human rights defenders, including those advocating for democracy and fair governance. While not focused on democracy per se, PBI helps safeguard activists, journalists, and civil society organizations working to uphold democratic and human rights. They typically have worked in countries with the same level of economic or military development as the US. The intervention arguably relies on local governments being (at least somewhat) intimidated by the involvement of people from outside the country, which may limit the intervention’s effectiveness. Full disclosure: I was previously on the board of a PBI entity.
Responsive (lawsuits, media, citizen action)
Lawsuits
I understand that there are several lawsuits currently underway for actions like the drastic evisceration of USAID, the attack on DEI, and the treatment of migrants. I’m confident that these will have some effect on maintaining the rule of law, however I’m not clear on how effective they will be. Can the President just ignore them?
Media
Trump is very negative towards certain areas of the media, and if he were to damage the freedom of the press in the US, this could limit Americans’ ability (and the world’s ability) to understand what’s happening in the US.
Citizen action
Citizens can certainly at least write to their representatives to express their concerns. Stronger actions (strikes and civil disobedience) should at some point be considered.
Activity
Brainstormed (ie not at all carefully researched) NGOs/orgs working on this
Lawsuits
American Civil Liberties Union (ACLU): The ACLU filed a lawsuit against the Trump administration, seeking access for attorneys to approximately 80 migrants transported from the U.S. mainland to Guantanamo Bay Naval Base. The suit argues that the government is denying the detainees their constitutional rights and calls for immediate legal access. wsj.com
Public Citizen: This advocacy group filed a lawsuit aiming to halt the administration's freeze on foreign aid, arguing that the suspension of congressionally allocated funds is illegal and endangers lives abroad. politico.com
Democracy Forward Foundation: This organization has filed multiple lawsuits opposing executive orders that significantly reduce USAID operations, implement a federal funding freeze, and alter federal employment policies. They argue these actions violate constitutional and administrative laws. en.wikipedia.org
National Association of Diversity Officers in Higher Education (NADOHE): NADOHE, along with other plaintiffs, filed a lawsuit seeking to block executive orders they claim unlawfully restrict diversity, equity, inclusion, and accessibility (DEIA) initiatives in higher education and other sectors. aalrr.com
Coalition of Christian and Jewish Denominations: Over two dozen religious groups sued the Department of Homeland Security, challenging the decision to permit Immigration and Customs Enforcement (ICE) agents to conduct raids in houses of worship, alleging it violates their First Amendment rights. en.wikipedia.org
Media
There may be scope to donate to support media sources which are willing to be critical of Trump. I understand these to include CNN, NPR, MSNBC.
Citizen action
There are several organisations for which you could volunteer and/or donate. The three below are simply the first three that I encountered.
Indivisible A grassroots network known for its progressive advocacy, Indivisible organizes local actions, including letter-writing campaigns and town hall meetings, to pressure elected officials on issues like healthcare, defense policy, and overall government accountability. Website:indivisible.org
MoveOn A progressive advocacy group that mobilizes millions of members through petitions, email campaigns, and coordinated actions to oppose policies they view as harmful—such as reductions in aid or questionable appointments—and to promote progressive reforms. Website: moveon.org
People for the American Way An advocacy organization that defends constitutional rights and democratic values. They coordinate campaigns and provide tools for citizens to contact their representatives about issues including executive overreach and poor governance. Website:pfaw.org
Preventative (free and fair elections)
Preventative actions include:
Preventing Trump from remaining in power for a third term: I’ll let other people who are better positioned than me comment on this.
Ensuring free and fair elections: this could involve lobbying against laws which limit access to voting, and strengthening voting rights laws, strengthening independent oversight such as nonpartisan election commissions, ensuring that election officials are protected, and supporting grassroots groups which increase voter participation are supported.
Activity
Brainstormed (ie not at all carefully researched) NGOs/orgs working on this
Specifically preventing Trump from coming to power for a third term
I’ll leave others to propose ideas, and act on them if they see fit.
Ensuring free and fair elections
Brennan Center for Justice – A legal think tank at NYU Law School that researches and advocates for policies to ensure voting rights, fair redistricting, and strong election security.
Common Cause – Works to reduce the influence of money in politics, prevent gerrymandering, and protect voting rights through advocacy and litigation.
Verified Voting – Focuses on election security, advocating for secure voting systems and verifiable paper ballots.
League of Women Voters (LWV) – A nonpartisan organization that promotes voter education, voter registration, and fair redistricting.
The Carter Center – Founded by former U.S. President Jimmy Carter, this organization monitors elections worldwide and provides technical assistance to ensure fair electoral processes. In recent years it expanded its work to cover US elections as well as international ones.
Below, I list some organisations that I left out of the table in the “free and fair elections” row.
Deprioritised because they don’t cover the US, as far as I know
Democracy International – Works on democracy and governance programs worldwide, including election monitoring, civic participation, and government accountability.
Transparency International – primarily focused on combating corruption, this NGO also addresses electoral fraud and promotes transparency in political funding.
FairVote – Advocates for electoral reforms in the U.S., such as ranked-choice voting, proportional representation, and voting rights protections. Possibly unfair to leave this off the main list in the table; I haven’t taken the time to understand how effective this work will be at reducing authoritarianism.
N of ITN: is there already enough happening to tackle this issue?
Alluded to earlier, the Eric Adams case demonstrates that some actions are being taken. However the extent of the response appears to have been tempered by the “firehose effect”. Ie there have been so many shocking events happening in quick succession that it’s hard for any one of them to elicit a meaningful response from most people. This may explain why the extent of the response seems low compared to the extent of the problem.
Is this only a matter for Americans?
Some might argue that this is a matter solely for Americans. However a quick glance at the list above (see section on I of ITN) indicates that most of the biggest risks that I’m worried about are global in nature, and therefore likely to ultimately affect all of us.
Nonetheless, I think it’s useful to draw a distinction between the party political (eg Trump and Harris may differ on policies, eg on attitudes to immigration) and general good governance. As I see it, it’s reasonable for anyone to promote general good governance, including free and fair elections and human rights. It’s also reasonable for this to be promoted regardless of borders. Interfering in the party political outcomes of another country is much more questionable.
Clearly, in this case, that distinction is more complex, and in this document I’m treading that line as carefully as I can.
Appendix: signs of democracy being under threat in the US
I’ll put these signs into the following categories:
Erosion of democratic institutions
Free and fair elections
Disinformation and misinformation
Attacks on the free press
Attacks on civil society
“Capture” of the private sector
Erosion of democratic institutions
Choices of leaders of departments, such as the department of health, or defence. Some appointments appear to be of individuals who do not have the relevant expertise and qualifications to perform the role effectively, which may undermine the effectiveness of the post-holder. Aside from the obvious issues that those departments run less effectively, such an individual may be a less effective check/balance against the authoritarianism of the President.
Trump wants to fire the director of the Office of Government Ethics (OGE) which monitors conflicts of interests. (Source: MSNBC)
Trump fired 17 inspectors general, whose role is to audit the actions of government. “'It's a widespread massacre. Whoever Trump puts in now will be viewed as loyalists, and that undermines the entire system,' one of the unnamed fired agents told the Washington Post. ” (copied from the Daily Mail) More detail in this article from Campaign Legal.
The judiciary: Trump has installed several judges who are allied to him, and those are lifelong appointments, which means long term support. The judiciary is an important component of the checks and balances which control the power of the president. (I haven’t looked into this carefully, and don’t understand the extent to which this departs from norms – every president tends to nominate judges who reflect their understanding of the law.)
Free and fair elections
Attempts to undermine the outcomes of the 2020 election
It is my understanding that Trump did not respect the outcome of the 2020 election, which he lost. I haven’t taken the time to research this position carefully. Some evidence which supports this position:
Pressuring State Officials: One widely reported incident involves Trump’s call to Georgia Secretary of State Brad Raffensperger, during which he urged him to “find” enough votes to overturn the election results in that state. This call is frequently cited as an effort to influence the outcome. (Sources: 1,2,3)
Legal Challenges and Claims of Fraud: Trump and his allies filed numerous lawsuits challenging the election outcome in several key states. These legal actions were consistently dismissed by courts for lacking credible evidence, yet they helped sow doubt among supporters about the legitimacy of the election. (Sources: 1,2,3)
Rhetoric and Public Statements: In the months following the election, Trump repeatedly made public statements alleging widespread voter fraud and a “stolen” election. These assertions, despite being debunked by multiple independent investigations and recounts, contributed to a broader narrative that sought to delegitimise the official results.
These actions, taken together, illustrate efforts that many observers and legal experts argue were intended to cast doubt on the legitimacy of the 2020 election results.
Treatment of 6 Jan rioters/”hostages”
Trump has pardoned the people who committed political violence on 6 Jan 2021, and who tried to overturn an election. (Sources: CNN, Reuters, Texas tribune). Respecting the outcome of free and fair elections is a cornerstone of a well governed democracy.
Voter suppression
Recent actions by the Trump administration have raised concerns about potential voter suppression. Notable examples include:
Implementation of the SAVE Act: The proposed Secure America's Vote and Elections (SAVE) Act mandates in-person proof of citizenship for voter registration. Critics argue this could disenfranchise millions, particularly married women who have changed their surnames, as well as people of color, young voters, and transgender individuals, due to documentation discrepancies.
Dismantling Election Security Measures: The administration has disbanded federal efforts to monitor election interference, including the dissolution of an FBI task force focused on foreign influence and significant staffing cuts at the Cybersecurity and Infrastructure Security Agency (CISA). These actions may increase vulnerabilities to foreign meddling in U.S. elections. This item overlaps with the point around disinformation and misinformation (see next section).
Reshaping the Justice Department's Role in Voting Rights: Plans to reduce enforcement of the Voting Rights Act and potentially prosecute local election officials have been reported. Such measures could undermine federal voting rights protections and intimidate officials responsible for overseeing elections.
These actions collectively suggest a pattern that may restrict voter participation and compromise the integrity of the electoral process.
Disinformation and misinformation
The global engagement centre has been deprived of funding, and it was the state department’s centre for fighting global disinformation. (source: Cyberscoop)
The Department of Homeland Security said on 11 Feb that personnel focused on misinformation, disinformation and foreign influence operations aimed at U.S. elections have been placed on administrative leave. (source: Reuters)
It’s been suggested that the USAID cuts will also exacerbate the misinformation problem (source: Guardian)
Attacks on the free press
Trump’s ability to reach followers directly, previously via twitter, now increasingly via truth social, arguably reduces his need to control the free press, even though throttling the free press has certainly been a core part of the authoritarian playbook in the past. Nonetheless, Trump has long been critical of the press, so some sort of hostile action is to be expected, and does, indeed, appear to have happened:
Trump has called for CBS to lose its licence, sued ABC news (source: Al jazeera)
Trump's administration also denied Associated Press access to the White House for failing to refer to the Gulf of Mexico as the Gulf of America (source: NBC)
Attacks on civil society
Recent actions by the Trump administration seem to suggest an undermining of civil society organisations, particularly non-governmental organizations (NGOs) and nonprofits. These measures include funding freezes, policy changes, and executive orders that collectively pose significant challenges to the operational capacity of these organisations.
For this section, the elements which I think are most relevant are attacks by the administration on those parts of civil society which are likely to hold government to account for their actions. I have not found evidence of this, although I have found plenty of examples of the administration attacking other parts of civil society, notably the USAID funding suspension, and an executive order which targets NGO funding. The latter may end up targeting the parts of civil society which I'm particularly considering here, but I don't know its outcome yet.
“Capture” of the private sector
The private sector can at times be a control over a government which is abusing its powers.
However recent actions by the Trump administration indicate a significant influence over private sector behavior, particularly through executive orders and policy directives that have prompted companies to modify their internal practices.
Impact on Diversity, Equity, and Inclusion (DEI) Initiatives:
On January 22, 2025, President Trump issued an executive order directing federal agencies to terminate DEI programs and to address "illegal private-sector DEI preferences, mandates, policies, programs, and activities." This directive has led several major corporations to reassess and, in some cases, scale back their DEI commitments.
Goldman Sachs: The financial institution has abandoned its policy of refusing to handle IPOs for companies with all-white, all-male boards, a reversal influenced by the administration's crackdown on DEI practices in the private sector. nypost.com
Hollywood Studios: Several media companies are reevaluating their DEI policies due to increased scrutiny from the administration, with some opting to quietly continue diversity efforts without explicitly labeling them as DEI initiatives. ft.com
Meta: Meta has decided to roll back its DEI programme.
Influence on Environmental, Social, and Governance (ESG) Investing:
The administration's stance has also impacted ESG investing, with policies that may deter companies from pursuing ESG-related initiatives.
Regulatory Changes: The administration is expected to challenge the Securities and Exchange Commission's rules requiring companies to disclose their emissions, potentially hindering transparency efforts related to environmental impact. internationalbanker.com
Corporate Strategy Adjustments: Companies may face increased pressure to align with the administration's policies, potentially leading to a reduction in ESG-focused investments and initiatives. velaw.com
Importantly, some of these changes are self-imposed, indicating that some corporates are already adjusting themselves to align with the Trump regime, without needing to be prompted. This raises concerns about the private sector's ability to rein in the administration.
48
More posts like this
151
Prediction
markets & many experts think authoritarian capture of the US looks distinctly possible
Looking back, it seems quite possible to me that EAs undervalued the importance of helping with the previous election. For one limited thing, having this administration be in power when we're getting so close to TAI seems like a major failure.
Personally, I think I've become more convinced recently that generic US policy, especially focusing on long-term issues (like US governance, or US decisions on questions like Nuclear/bio/AI) might be a good use of EA funds.
I would have really guessed that a lot of this area wouldn't at all be neglected, but in practice, it seems to be far more ignored than I think is reasonable.
I agree that US policy is obviously very important, but
I think it always has been? My sense is that lots of EA funds are already spent on US policy things, e.g. https://www.nti.org/analysis/ and https://www.governance.ai/research
SBF was willing to bribe Trump w $5bi so he wouldn't run in 2024. That's not what I'd call "undervalue".
I think SBF was fairly unique in caring about this, and his empire collapsed before he did anything like this in that election. When I said "undervalue", I wasn't referring to SBF, given he hasn't been active in this time. (Obvious flag that while I might have sympathized with some of SBF's positions, I very much disagree with many of his illegal and fraudulent actions)
Looking back though, paying $5B for Trump to definitely not run/win seems like a great deal to me, though the act of paying him raises a lot of qualms I'd be uncomfortable with.
And we should also examine neglect not just on the headline number of dollars going into the space but on specific facets, like how much money in that space actually goes to top tier impact opportunities or how much investment is there in innovating the space/interventions.
I think in-comparison to the space, EA has a comparative advantage more in talent than in money. I think the Harris campaign got $2B or so of donations, but I get the impression that it could have used smarter + more empirically-minded people. That said, there is of course the challenge or actually getting those people to be listened to.
I have not seen a lot of evidence that EA skills are very transferable to the realm of politics. As counterexamples, look at the botched Altman ouster, or the fact that AI safety people ended up helping start an AI arms race: these partially seem to come from a place of poor political instincts. EA is also disproportionately STEM background, which are generally considered comparatively poor at people skills (accurately, in my experience).
I think combating authoritarianism is important, but EA would probably be better off identifying other people who are good at politics and sending support their way.
I disagree, but this has me curious.
My impression from other writing I've seen of yours is that you don't think that EAs are good at too many things. What do you think EAs are best at, and/or should be doing? Perhaps, narrow GiveWell-style research on domains with lots of data?
I think you probably need multiple kinds of skill and some level of cognitive style diversity within a political campaign. You definitely need a lot of people with people skills, and I am sure that the first gut instincts of people with good social skills about what messaging will work are better than those of people with worse social skills. Those socially skilled people should undoubtedly be doing detailed messaging and networking for the campaign. But you also need people who are prepared to tell campaigns things they don't want to hear, even when there is severe social pressure not to, including things about what data (rather than gut instinct) actually shows about public opinion and messaging. (Yes, it is possible to overrate such data which will no doubt be misleading in various ways, but it also possible to underrate it.) My guess is that "prepared to tell people really hard truths" is at least somewhat anticorrelated with people skills and somewhat correlated with STEM background. (There is of course a trade-off where the people most prepared to tell hard truths are probably less good at selling those truths than more socially agreeable people.) For what it's worth Matt Yglesias' seems pretty similar to the median EA in personality, and I recall reading that Biden advisors did read his blog. Ezra Klein also seems like a genuinely politically influential figure who is fairly EA-ish. There is more than one way to contribute to a political movement.
I personally don't think EA should be doing much to combat authoritarianism (other than ideally stopping its occasional minor contributions to it via the right-wing of rationalism and being clear-eyed about what the 2nd Trump admin might mean for things like "we want democratic countries to beat China") because I don't think it is particularly tractable or neglected. But I don't think it is a skill issue, unless you're talking about completely EA run projects (and even then, you don't necessarily have to put the median EA in charge; presumably some EAs have above average social skills.)
Actually I think this is the one thing that EAs could realistically do as their comparative advantage, considering who they are socially and ideologically adjacent to, if they are afraid of AGI being reached under an illiberal, anti-secular, and anti-cosmopolitan administration: to be blunt, press Karnofsky and Amodei to shut up about "entente" and "realism" and cut ties with Thiel-aligned national security state companies like Palantir.
I don't think cutting ties with Palantir would move the date of AGI much, and I doubt it is the key point of leverage for whether the US becomes a soft dictatorship under Trump. As for the other stuff, people could certainly try, but I think it is probably unlikely to succeed, since it basically requires getting the people who run Anthropic to act against the very clear interests of Anthropic and the people who run it (And I doubt Amodei in particular, sees himself as accountable to the EA community in any way whatsoever.)
For what it's worth I also think this complicated territory and that there is genuinely a risk of very bad outcomes from China winning an AI race too, and that the US might recover relatively quickly from its current disaster. I expect the US to remain somewhat less dictatorial than China even in the worst outcomes, though it is also true that even the democratic US has generally been a lot more keen to intervene, often but not always to bad effect, in other country's business.
Conditional on AGI happening under this administration, how much AGI companies have embedded with the national security state is a crux for the future of the lightcone, and I don't expect institutional inertia (the reasons why one would expect "the US might recover relatively quickly from its current disaster" and "the US to remain somewhat less dictatorial than China even in the worst outcomes") to hold if AGI dictatorship is a possibility for the powers that be to reach for.
"how much AGI companies have embedded with the national security state is a crux for the future of the lightcone"
What's the line of thought here?
It increases the AI arms race thus shortening AGI timelines, and, after AGI, increases chances of the singleton being either unaligned or technically aligned to being an AGI dictatorship or other kind of dystopian outcome.
Thinking about this a bit more -
My knee-jerk reaction is to feel attacked by this comment, on behalf of the EA community.
I assume that one thing that might be going on is a miscommunication. Perhaps you believe that I was assuming that EAs could quickly swoop in, spent a little time on things, and be far more correct than many experience political experts and analysts.
I'm not sure if this helps, but the above really doesn't align with what I'm thinking. More something like, "We could provide more sustained help through a variety of methods. People can be useful for many things, like direct volunteering, working in think tanks, being candidates, helping prioritization, etc. I don't expect miracle results - I instead expect roughly the results of adding some pretty smart and hardworking people."
The usefulness of smart people is highly dependent on the willingness of the powers-that-be to listen to them. I don't think lack of raw intelligence had much of anything to do with the recent US electoral results. The initial candidate at the top of the ticket was not fit for a second term, and was forced out too late for a viable replacement to emerge. Instead, we got someone who had never polled well. I also don't think intelligence was the limiting factor in the Democrats' refusal to move toward the center on issues that were costing them votes in the swing states. Intellectually understanding that it is necessary to throw some of your most loyal supporters under the bus is one thing; committing to do it is something else; and actually getting it done is harder still. One could think of intelligence as a rate-limiting catalyst up to a certain point, but dumping even more catalyst in after that point doesn't speed the reaction much.
I think @titotal's critique largely holds if one models EAs as a group as exceptional in intelligence but roughly at population baseline for more critical and/or rate-limiting elements for political success (e.g., charisma, people savvy). I don't think that would be an attack -- most people are in fact broadly average, and average people would be expected to fail against Altman, etc. And if intelligence were mostly neutralized by the powers-that-be not listening to it, having a few hundred FTEs (i.e., ~10% of all EA FTEs?) with a roughly normal distribution of key attributes is relatively unlikely to be impactful.
Finally, I think this is a place where EA's tendencies toward being a monoculture hurts -- for example, I think a movement that is very disproportionately educationally-privileged, white, STEM focused, and socially liberal will have a hard time understanding why (e.g.) so many Latino voters [most of whom share few of those characteristics] were going for Trump this cycle and how to stop that.
On EAs in policy, I'd flag that:
- There's a good number of people currently working in AI governance, Bio governance, and animal law.
- Very arguably, said people have had a decent list of accomplishments and power positions, given that such work was fairly recent. See Biden's executive orders on AI, or the UK AI Security Institute. https://www.aisi.gov.uk/
- People like Dustin Moskovitz and SBF were some highly prominent donors to the Democratic party.
I think the EA policy side might not get a huge amount of popularity here, but it seems decently reputable to me. Mistakes have been made, but I think a decent report on the wins and losses would include several wins.
I do agree that finding others doing well and helping them is one important way to help. I'd suspect that the most obvious EA work would look like prioritization for policy efforts. This has been done before, and there's a great deal more that could be done here.
In fairness, SBF was also secretly a prominent Republican donor, right? Didn't he basically suggest in the infamous interview with Kelsey Piper that he was essentially cynical about politics and just trying to gain influence with both parties to help advance FTX and Alameda's interests?
He was a Republican donor, but from what I understand, not really a MAGA donor. My impression was that he was funding people on both sides, who were generally in favor of their interests - but that their interests did genuinely include issues like bio/ai safety.
I think it's very reasonable to try to be bipartisan on these issues.
Fair point. I certainly don't think it is established (or even more than 50% likely) that SBF was purely motivated by narrow personal gain to the exclusion of any real utilitarian convictions at all. But I do think he misrepresented his political convictions.
I think it's clear he misrepresented his political convictions, especially to the public (as opposed to close friends and some EAs).
But I think there's separately decent evidence that he was thinking of himself as ultimately advancing utilitarian goals.
Not that that makes it okay - it's very possible to consider yourself trying to help any noble goals - then using that to justify really bad actions.
That's my read of the evidence as well, but I haven't examined it closely.
What about this current administration and possible TAI is the failure? Is orthogonal to American 'democracy' as outlined in the post?
It is orthogonal. More that TAI might be soon, we probably want an administration that would both promote AI safety and broadly be cooperative/humble/deliberate.