I overall like this post and I find the arguments that electing Democrats is really important and donating to strong candidates helps achieve that goal convincing, but it doesn’t really make a case for those donations being more cost-effective than donating to AI safety orgs or GiveWell. Do you know of any analysis which tries to directly compare the expected effects of spending a dollar on the most important elections to spending it on e.g. malaria nets? Or do you have any arguments for why we should expect political donations to have competitive cost-effectiveness without that?
The AI safety people I know mostly believe that donating to political candidates is at current margins better than donating to AI safety organizations because the lowest-hanging fruit for AI safety organizations has generally been picked by Coefficient Giving, the Survival and Flourishing Fund, etc. (If you have idiosyncratic views of how AI safety works, that might not be true.)
I don't think it's possible to directly compare without making up numbers in a way that I think is more confusing than enlightening. But you can yourself make up numbers about how much you think it might cost to flip a vote, the chance that the Democrats would bring back USAID, and the disvalue associated with America becoming an illiberal autocracy. They're probably as good as anyone else's.
I agree with this, though there are some people who I think have tried (or are trying) to put some more effort into thinking this through (with the same conclusion).
The hard money cap creating a more level playing field is underrated. When someone like Moskovitz can only donate the same $7k as anyone else making 99th percentile income, it fundamentaly shifts how political influence works. I worked on a few local campaigns and we always prioritized individual donors over PAC money becuase it showed genuine grassroots support. That constraint forces a broader coalition instead of dependency on mega-donors.
Any political analysis that says "politician X is universally bad" is pretty much guaranteed to be poorly thought out and driven by some kind of hysteria rather than rational thinking.
Trump 2 is so far looking very good for American democracy. Lets go through some of your claims:
> Trump is taking control of traditionally independent positions in the civil service—from prosecutors to the head of the Bureau of Labor Statistics—and trying to fill them with political loyalists who will do his bidding.
Think for a moment about what you are saying here. I remember a time when Republicans thought President Biden was going senile and that the deep state was secretly running everything behind the scenes. The mainstream narrative at the time tried to frame these Republicans as crazy. Then it became undeniable that President Biden was going senile, and we never got a straight answer about who had been running things behind the scenes. Now you are telling us that of course the deep state is running everything behind the scenes, and that it is somehow a threat to democracy for the democratically elected leader of our government to try to run things instead?!?!
This I think is our biggest crux. The idea that any bureaucrat can exercise part of the awesome power of the government without political accountability is horrifying to me. The narrative I was given at every level of our education system, from elementary school through law school, was that it was ok for prosecutors and bureaucrats to exercise that power, because they all answer to the president, who in turn answers to the voters. That was what I read in the United States Constitution when it said that all of the executive power is vested in the president. That was the thing that was supposed to safeguard us from bureaucrats and prosecutors becoming dictators. Now we are being told that that isn't true, that that was never true, that these prosecutors and bureaucrats have been acting as dictators all along. I want my democracy back! And Trump is fighting to get me my democracy back. In the long run this may do more to strengthen democracy than his attempted coup did to damage it.
> He has used retaliatory arrests and prosecutions against people who investigated him.
There has definitely been damage to democracy here, I grant, but it was the Democrats who started it, not Trump. You remember the prosecution of Trump over the Stormy Daniels thing? A prosecution that in normal times nobody would have bothered with even if it had been legal, and that was definitely not legal because it violated the statute of limitations. The Democrats crossed a huge red line there and took our democracy in a very dark direction. I wish we had a president who could pull us back from that, and Trump sadly but unsurprisingly is not that man, but I also can't really blame him for playing by the same rules that were used against him.
> He has openly defied as many as one in three court orders against him.
I don't believe this for a second. The first couple times I saw these claims I looked into them and they did not hold up. I'm no longer taking the time to look in to each individual case, and I certainly am not going to give the Washington Post my money, but if you have a particular case in mind that you think is a particularly strong example of this I am willing to look into it.
> He has broken the law to avoid spending money allocated by Congress.
Yea, that's not great. You know what else is not great? His predecessor breaking the law to avoid collecting debts that former college students owed to the American people. Again, this just looks like Trump playing by the same rules the Democrats played by the last time they were in power.
> He has taken power from the states—most famously by deploying California’s National Guard without the consent of the governor of California.
The constitution allocates power over the military, including the national guard, to the President, not to the governors. And it is sometimes right and proper for a president to use that power to keep the peace and enforce the law. I recall another Republican president famously deploying the 101st Airborn Division to Little Rock, AK, a decision now rightly celebrated in the history books. It may have been unwise or even illegal for Trump to deploy the guard under the current circumstances, but it was definitely not an invasion of state power.
> He has also targeted the institutions of civil society that are supposed to check the power of the president: the press, law firms, universities.
He has been more political than some other presidents in choosing which elements of the press to give access to, but access has always been the president's to give. Biden took advantage of this to hide from the press when he was going senile, and nobody said he was targeting the press.
It is the sacred right of every litigant to choose their own lawyers, and fire their own lawyers, for whatever petty and childish reasons they choose. As head of the executive branch, it is Trump's sacred right to choose the executive branches lawyers, and fire the executive branches lawyers, for whatever petty and childish reasons he chooses. That's all he's done.
The universities were NEVER supposed to be a check on the power of the president. That is a horrifying misuse of those institutions, and the reason the president is rightly coming down on them. Universities are supposed to pursue truth outside the realm of politics, and they are supposed to do so in a way that is welcoming to people of all political persuasions, all races, all sexes, and all religions, and that does not take sides on political issues or politicians. That was the promise that universities made to the American people in exchange for the incredible degree of financial support and independence they have received since ww2. The Universities broke that contract. And that sucks. I really wish they hadn't, because they were important institutions, and they have left a vacancy that will be difficult to fill. But what is happening now is the natural and foreseeable consequence of that choice.
> Most frighteningly, Trump has continued to say he won the 2020 election. He pardoned the January 6 rioters who, again, tried to perform a coup against the U.S. government to make him president. He has targeted the people who prosecuted the January 6 rioters. He has hinted publicly at wanting to run for an unconstitutional third term.
Yea, you are definitely right that in past generations attempting a coup would have been disqualifying. I wish it still was. This is definitely the thing that gives me most pause about Trump, and the dominant reason I voted against him in 2024. But the American people have spoken, and they clearly do not regard it as disqualifying. As for a third term, I'd be shocked if it gets any real traction. And if, somehow, it does happen, I'll join the resistance then.
If there is an EA case against Trump, I think it would have to be that all this democracy stuff is too speculative and uncertain to satisfy the rigorous epistemological standards we demand of ourselves as EAs, and the damage Trump has done on foreign aid and foreign policy more generally is worth fighting against. There is a whole argument that by pulling the US back from its traditional role as world cop, Trump has significantly increased the risk of ww3 breaking out, and the catastrophic and existential risks that go along with that. You should be making that argument!
A lot of this is true. You might look at, say, selective prosecutions of companies for discriminating against blacks and women but not against whites and men and figure 'sauce for the goose, sauce for the gander'. And as you point out Biden's revocation of student debt was illegal too. And I agree the universities have become mouthpieces of the left, destroying their wider credibility, and that's part of the reason we're seeing such a huge right-wing reaction now.
However, we're seeing things like the Bureau of Labor Statistics stopping collecting data, the CDC coming out against vaccines, and giant tariffs and actions that enrage longtime allies like Europe and Canada. This is doing serious damage to state capacity, and as a 'liberal nationalist', to quote Eric Kauffman, this is bad for the country long-term. Overall increases in the level of corruption will take a long time to undo, and I don't think we'll ever get as close to Canada as we were pre-Trump.
I definitely agree they crossed the line going after him over Stormy Daniels. I don't really care about that. However, he did try to launch a coup on January 6, and I wish they'd stuck to going after him on that.
I think the USA should stay a democracy. We weren't perfect, but we were reasonably close to the democratic end of the spectrum throughout our history, and we're likely to get incompetent kleptocrats if we go autocratic. The Chinese have a long tradition of competent bureaucracy checking and moderating the emperor and of the duties of ministers. It's not perfect either, but there's something recognizably Chinese after 2000 years and 7+ dynasties. I doubt we will be so lucky. As for the Russians, well, they've long recognized whatever their nation's strengths it's always been poorly run.
EA assumes all lives are of equal worth, so the damage caused by the end of foreign aid is more than enough from the EA point of view I think. (Personally I thought it was useful as an instrument of soft power.)
there is one point i argue about in context of Israel and now i extrapolate to USA:
There are two ways to understand the word "democracy". The Rule Of The People, and Balances And Brakes.
i think about it as Democracy and Liberalism but i don't think it's proper translation to English.
there are people who see Democracy as the system of governance when people decide by voting, when people are the sovereign, and restriction on it is Undemocratic.
and there are people who see Democracy as system of governance that strongly restricting what the people can decide to do, when the restrictions are the heart of the system. you surely see a lot of this desire to restrict in USA governance!
for people in the second group, the fact that the president is not above the law is important part of democracy. rule that say, for example, execute all Jews, is Undemocratic, even if 80% of the people vote for it.
for the first grope, the decision of 80% of the people is Democracy, and preventing it is Undemocratic.
and instead of talking about this important disagreement, people don't notice it, because those two different principles referred by the same word, Democracy. or even worst, notice and then try to win the semantic game, as if it will let them win the actual game.
I think you are being too charitable. I think it is true that there are sometimes conflicts between democracy and other values. Some limitations on democracy are good. That is the point of our bill of rights and our equal protection clause. Killing all the Jews would surely be bad, even if democracy were to demand it.
Where I think we disagree is that I don't see the ambiguity in the English language that you do. "Democracy", as I see it, unambiguously refers to rule by the people. Before Trump, and to a large extent even just before Trump 2, that was how it was used. We are now in a world where a lot of liberals, unable to accept that Trump won a legitimate democratic election, twice, are now using the word "democracy" in a dishonest way - they have started using it to refer to "balances and breaks" as you call it, to restrictions. They are trying to challenge democracy because they don't like the leader democracy gave us, and they aren't willing to admit that fact. That is what is going on here. And I will not play into it.
Your problem is that you have bought into the simplistic argument used by dictators the world over: "I won an election, therefore the democratic thing is for me to be able to do whatever I want." This is technically true for a certain definition of democracy, but tends to result in cataclysmically bad government. This also tends to be undemocratic in the sense that it results in outcomes disfavored by most people, like massive economic damage and restriction of freedom. The idea that democracy must be tempered by checks and balances is not new, indeed you may find the phrase used in the Federalist Paper titled “The Structure of the Government Must Furnish the Proper Checks and Balances Between the Different Departments.”
To pick one example of the principle, consider racial segregation. In your sense, Jim Crow laws were democratic in the sense that a majority of voters had voted for them at one point. In another sense, they were undemocratic because by suppressing the vote of a significant portion of the population they damaged the ability of the electoral system to accurately judge what the majority of people wanted. This is why we have a Constitution in the first place: so that one 51% election victory cannot overturn the foundations of society.
The American left is quite aware that Trump won the popular vote (this time at least). The most popular topic of inter-left discourse right now is "Why does everyone hate us more than the most contemptible man alive, is it the fault of the moderates or the progressives?" which very much accepts the turn of public opinion in its premise. The main linguistic compression here is that "democracy" is typically used to mean "liberal democracy", the version of democracy that's patched to avoid demagogue strongmen ruining everything, because that's the version this country has run on since its foundation. More detailed coverage of this topic can be found here: astralcodexten.com/p/defining-defending-democracy-contra
Don't put words in my mouth. I never said democracy was the only value or that Trump should be able to do anything he wants. I took issue with much more specific points that Ozy made. If you disagree with any of the points I made, feel free to respond to them.
I, personally, find Semantic Arguments as confusion at best, and attempts to Sneak In Connotations at worst. I even less equipped to talk about semantics in English, my third language.
but I have one friend that honestly see Checks And Balances as important part of the definition of Democracy, and i don't find it surprising, as i, too, learned that in school. you probably learned something else.
generally speaking, it's the Progressives who choose Sneaking In Connotation as major strategy, and if you will claim that people who believe that is instance of movement believing its own propaganda, you probably right.
and yet, the conclusion remain the same - you just waste time, arguing about semantics instead of content. instead of arguing what Democracy is, you can say that the people are the sovereign, and that the majority should rule, and not some aristocratic minority.
i'm all in favor of naming the thing, the more the better. what can be destroyed by naming it should. but the way to do it is by pointing the substance, not arguing semantics. there are already too much of the debate in the third level of simulacra.
I can't make US political donations, but consider adding Manny Rutinel, running for Colorado's 8th district, to the list? I once had a conversation with him about farmed animal welfare, he helped pass State AI law (not sure how good it was, but an indication that he cares https://x.com/MannyRutinel/status/1798086169290498395?s=20 ) and he's running in a purple district.
This is one of those times where I have to decide whether I'm an ex-liberal angry at feminism, or an ex-conservative angry at corruption.
Honestly, I'll probably do nothing because I'm afraid of retaliation by the administration and I'm a coward.
I want to say I really admire your focus on pragmatism over ideological purity. Yes, it is a good thing to work in the current administration if it means you can push back on antivax!
About the AI... this is one of those things where I don't understand the tech enough to have an opinion, and everyone who does either seems to have a political ax to grind or investments in the companies.
Let me see if I can make a mostly apolitical, non-technical argument in favor of being careful about AI.
PART 1 (assumption). We might be close to actually building something which makes humans economically and evolutionarily obsolete.
This is difficult to visualize or believe, because it has never happened before. But imagine if we *did* build something that could do any task that a Stanford graduate could do for $1/hour: Software engineering, hardware engineering, basic science, being a (good) CEO, operating a robot, designing a better robot, winning an election, lobbying for legislation, consolidating power. Basically, any human task, done well, for $1/hour. Really think about what this deeply strange world would look like.
(If you like, imagine the AIs themselves making decisions, or imagine small numbers of powerful human beings controlling the AIs. It makes less difference than you'd think.)
The AI safety argument rests on the idea that we are potentially "0-2 major breakthroughs" away from this, plus some engineering work. ("Major breakthrough" here means "a paper like 'Attention is All You Need'", which typically come along every 5-10 years.) Sometime between, say, 2028 and 2050, human labor would become absolutely, completely, permanently worthless. Humans would be as economically competitive as chimpanzees.
Will are the odds this will actually happen? Experts disagree. But the AI safety argument is based on the assumption that there's a significant chance this might be a real thing that actually happens within our lifetimes. If you don't buy this idea, then nothing about AI safety will actually make sense.
PART 2 (argument). If a powerful entity and a powerless entity want the same limited resources, the powerless entity is likely to have a bad day.
Evidence: History, economics, colonialism, war, politics, etc.
If it's possible to get any kind of skilled or unskilled labor for $1/hour, then ordinary humans are screwed. Maybe the AIs have their own goals. Maybe the AIs all obediently listen to Billionaire X or Politician Y. But no matter who is in charge, every ordinary human consuming resources is "wasting" resources that the AIs could use better to accomplish the goals of whoever is in charge. The ordinary humans have *nothing* to offer. And if the ordinary humans revolt, the AIs are *also* better at war.
Even if this situation works out for 5 years or 20 years, it's not stable.
So that's my argument: If you make the human mind economically obsolete in every important niche, then it's likely very hard to stop economics and evolution from doing their thing.
But if you don't accept PART 1 (human obsolescence) as a potential risk, then PART 2 (obsolete things tend to get brutally marginalized) doesn't follow.
(There are other possible takes, including "humans suck, let's take a chance on the AIs that replace us." I don't plan to address these right now.)
That's quite reasonable and I've often thought the same. I just feel like a lot hinges on Part 1, and we generally haven't seen the computers be quite as good at blue-collar or pink-collar (service) labor as they are at white-collar lately. Frankly I'm quite pessimistic about the future--I think the billionaires own everything and will just let us all starve.
And on rationalist or rationalist-adjacent blogs I generally assume almost everyone is better informed about this AI stuff than me, so I stick to other topics.
From the link: "Alex Bores worked at Palantir for a few years. I think that many of Palantir’s activities are pretty evil. That said, as far as I know, Bores did not work on any such projects while at Palantir, and he says that he left Palantir over concerns about its collaboration with ICE."
Working for Thiel for five years, until 20-goddamn-19, and then leaving because of some very vague discomfort over "collaboration with ICE," and *then* immediately running for office as a Democrat raises some obvious questions, such as: "Are you stupid?" and "Do you think I am stupid?" This is a gigantic scarlet letter and I would need to hear an answer, from him, that is specific and convincing. So would the average Manhattan primary voter.
I'm sure he speaks fluent quokkese, but in a seat as safely blue as this one I see little reason to risk a Tricia Cotham situation.
There are many things wrong with the political assumptions in this post, but thankfully the nonsense is clear in two consecutive paragraphs:
"Obviously, you shouldn’t take a position that makes the Trump administration more effective at overthrowing American democracy. But the first Trump administration was an ordinary Republican administration because a huge number of policymakers, advisors, and civil servants thwarted Trump’s authoritarian instincts"
You support unelected bureaucrats having power to maintain the policies they want regardless of the results of elections. But surely you support it because they would stand up for freeedom?
"keeping vaccines somewhat more accessible, making AI regulation somewhat saner, preserving whatever scraps of foreign aid they can"
So this is what you think oposing "authoritarianism" means : ensuring state intervention and regulation.
Your piece barely mentions the policies where the republicans do want intervention and regulation. You want progressives to maintain state control through unelected bureaucrats, and do so in the name of "liberal democracy"
The bit where people are opposing authoritarianism is the bit you left out! But foreign aid, vaccines, and AI regulation are also important issues.
I am baffled by the claim that it is state intervention and regulation to think the federal government should continue to approve vaccines. If we abolished the FDA, any company could put out whatever vaccines they want.
I didn't mention other stuff because I don't give a shit about it compared to democracy, USAID, and AI. For example, Republicans are anti-trans, but as a trans person if polling suggested that a candidate punching a trans person in the face would flip a House seat blue, I would be the first in line to get my teeth knocked out.
New EA aligned research org powerfordemocracies.org has a cost effectiveness breakdown based on ITN and recommends the non profit Freedom2Vote with $94 to increase a voter. IMHO that's more aligned to how EAs would think about impact on the margin.
Generally agree with this, and those of us who work in this space professionally (and think in this way) generally kind of maintain a "supply curve" of different fundable opportunities at different prices (which, of course, can change *within* each opportunity at different budgets). However, there is a massive and important universe of things that can't be measured - things that potentially are more effective than single tactics like Freedom2Vote which can be measured - and so we also shouldn't fall into the "looking under the lamppost because that's where the light is" problem.
Just as an obvious example, a large funder could be thinking right now about ways to increase the chances that the ‘28 Dem primary nominates the best and most electable candidate. You can’t run an RCT on that, but it plausibly dwarfs Freedom2Vote (and I’m a huge fan of Freedom2Vote!).
A better way to convince EA minded donors will be a prioritization framework comparing these candidates against others as well as Freedom2Vote with the funding in margin. I'm sure the industry has some of these supply curve data and analyses and if they're willing to share that could potentially convince more EA minded donors. But in general if you claim of the potential to have a bigger impact without a quant analysis of how much that could be, it's harder to convince EAs.
Orange man is obviously bad (anybody saying otherwise is a psycho or a moron), but he didn't get into the White House twice randomly. Both the Dem and Rep establishments (and their analogues in the rest of the west) are rotten to the core, in thrall to absurd bankrupt ideologies and mainly concerned with zero-sum status games. This state of affairs is unsustainable and deteriorating, and the likes of Trump getting elected is just a symptom. Getting a couple marginally more sane (even granting that they are indeed such) people in office won't change the overall trajectory of the civilization (at best slightly delaying the collapse), and EAs would do well not to fool themselves otherwise.
if you believe that AI may soon come, delaying civilization collapse is pretty important, actually. and even if not, delaying civilization collapse is much better then... what is the suggested alternative? do nothing and watch Rome burning?
>what is the suggested alternative? do nothing and watch Rome burning?
Well, the EA calling is to look for "important, tractable, neglected" interventions, and who's to say that none remain that would be enough to avert the collapse, instead of slightly delaying it?
I like this post, especially the sober Trump evaluation.
I do feel there's some validity to the idea that "There should be an AI risk evaluation committee" really means "I want to be paid to be on a committee and doomsay about AI." But it's a toss-up between who should be trusted less, them or the Tech-Billionaires.
I overall like this post and I find the arguments that electing Democrats is really important and donating to strong candidates helps achieve that goal convincing, but it doesn’t really make a case for those donations being more cost-effective than donating to AI safety orgs or GiveWell. Do you know of any analysis which tries to directly compare the expected effects of spending a dollar on the most important elections to spending it on e.g. malaria nets? Or do you have any arguments for why we should expect political donations to have competitive cost-effectiveness without that?
The AI safety people I know mostly believe that donating to political candidates is at current margins better than donating to AI safety organizations because the lowest-hanging fruit for AI safety organizations has generally been picked by Coefficient Giving, the Survival and Flourishing Fund, etc. (If you have idiosyncratic views of how AI safety works, that might not be true.)
I don't think it's possible to directly compare without making up numbers in a way that I think is more confusing than enlightening. But you can yourself make up numbers about how much you think it might cost to flip a vote, the chance that the Democrats would bring back USAID, and the disvalue associated with America becoming an illiberal autocracy. They're probably as good as anyone else's.
I agree with this, though there are some people who I think have tried (or are trying) to put some more effort into thinking this through (with the same conclusion).
The hard money cap creating a more level playing field is underrated. When someone like Moskovitz can only donate the same $7k as anyone else making 99th percentile income, it fundamentaly shifts how political influence works. I worked on a few local campaigns and we always prioritized individual donors over PAC money becuase it showed genuine grassroots support. That constraint forces a broader coalition instead of dependency on mega-donors.
Any political analysis that says "politician X is universally bad" is pretty much guaranteed to be poorly thought out and driven by some kind of hysteria rather than rational thinking.
Trump 2 is so far looking very good for American democracy. Lets go through some of your claims:
> Trump is taking control of traditionally independent positions in the civil service—from prosecutors to the head of the Bureau of Labor Statistics—and trying to fill them with political loyalists who will do his bidding.
Think for a moment about what you are saying here. I remember a time when Republicans thought President Biden was going senile and that the deep state was secretly running everything behind the scenes. The mainstream narrative at the time tried to frame these Republicans as crazy. Then it became undeniable that President Biden was going senile, and we never got a straight answer about who had been running things behind the scenes. Now you are telling us that of course the deep state is running everything behind the scenes, and that it is somehow a threat to democracy for the democratically elected leader of our government to try to run things instead?!?!
This I think is our biggest crux. The idea that any bureaucrat can exercise part of the awesome power of the government without political accountability is horrifying to me. The narrative I was given at every level of our education system, from elementary school through law school, was that it was ok for prosecutors and bureaucrats to exercise that power, because they all answer to the president, who in turn answers to the voters. That was what I read in the United States Constitution when it said that all of the executive power is vested in the president. That was the thing that was supposed to safeguard us from bureaucrats and prosecutors becoming dictators. Now we are being told that that isn't true, that that was never true, that these prosecutors and bureaucrats have been acting as dictators all along. I want my democracy back! And Trump is fighting to get me my democracy back. In the long run this may do more to strengthen democracy than his attempted coup did to damage it.
> He has used retaliatory arrests and prosecutions against people who investigated him.
There has definitely been damage to democracy here, I grant, but it was the Democrats who started it, not Trump. You remember the prosecution of Trump over the Stormy Daniels thing? A prosecution that in normal times nobody would have bothered with even if it had been legal, and that was definitely not legal because it violated the statute of limitations. The Democrats crossed a huge red line there and took our democracy in a very dark direction. I wish we had a president who could pull us back from that, and Trump sadly but unsurprisingly is not that man, but I also can't really blame him for playing by the same rules that were used against him.
> He has openly defied as many as one in three court orders against him.
I don't believe this for a second. The first couple times I saw these claims I looked into them and they did not hold up. I'm no longer taking the time to look in to each individual case, and I certainly am not going to give the Washington Post my money, but if you have a particular case in mind that you think is a particularly strong example of this I am willing to look into it.
> He has broken the law to avoid spending money allocated by Congress.
Yea, that's not great. You know what else is not great? His predecessor breaking the law to avoid collecting debts that former college students owed to the American people. Again, this just looks like Trump playing by the same rules the Democrats played by the last time they were in power.
> He has taken power from the states—most famously by deploying California’s National Guard without the consent of the governor of California.
The constitution allocates power over the military, including the national guard, to the President, not to the governors. And it is sometimes right and proper for a president to use that power to keep the peace and enforce the law. I recall another Republican president famously deploying the 101st Airborn Division to Little Rock, AK, a decision now rightly celebrated in the history books. It may have been unwise or even illegal for Trump to deploy the guard under the current circumstances, but it was definitely not an invasion of state power.
> He has also targeted the institutions of civil society that are supposed to check the power of the president: the press, law firms, universities.
He has been more political than some other presidents in choosing which elements of the press to give access to, but access has always been the president's to give. Biden took advantage of this to hide from the press when he was going senile, and nobody said he was targeting the press.
It is the sacred right of every litigant to choose their own lawyers, and fire their own lawyers, for whatever petty and childish reasons they choose. As head of the executive branch, it is Trump's sacred right to choose the executive branches lawyers, and fire the executive branches lawyers, for whatever petty and childish reasons he chooses. That's all he's done.
The universities were NEVER supposed to be a check on the power of the president. That is a horrifying misuse of those institutions, and the reason the president is rightly coming down on them. Universities are supposed to pursue truth outside the realm of politics, and they are supposed to do so in a way that is welcoming to people of all political persuasions, all races, all sexes, and all religions, and that does not take sides on political issues or politicians. That was the promise that universities made to the American people in exchange for the incredible degree of financial support and independence they have received since ww2. The Universities broke that contract. And that sucks. I really wish they hadn't, because they were important institutions, and they have left a vacancy that will be difficult to fill. But what is happening now is the natural and foreseeable consequence of that choice.
> Most frighteningly, Trump has continued to say he won the 2020 election. He pardoned the January 6 rioters who, again, tried to perform a coup against the U.S. government to make him president. He has targeted the people who prosecuted the January 6 rioters. He has hinted publicly at wanting to run for an unconstitutional third term.
Yea, you are definitely right that in past generations attempting a coup would have been disqualifying. I wish it still was. This is definitely the thing that gives me most pause about Trump, and the dominant reason I voted against him in 2024. But the American people have spoken, and they clearly do not regard it as disqualifying. As for a third term, I'd be shocked if it gets any real traction. And if, somehow, it does happen, I'll join the resistance then.
If there is an EA case against Trump, I think it would have to be that all this democracy stuff is too speculative and uncertain to satisfy the rigorous epistemological standards we demand of ourselves as EAs, and the damage Trump has done on foreign aid and foreign policy more generally is worth fighting against. There is a whole argument that by pulling the US back from its traditional role as world cop, Trump has significantly increased the risk of ww3 breaking out, and the catastrophic and existential risks that go along with that. You should be making that argument!
A lot of this is true. You might look at, say, selective prosecutions of companies for discriminating against blacks and women but not against whites and men and figure 'sauce for the goose, sauce for the gander'. And as you point out Biden's revocation of student debt was illegal too. And I agree the universities have become mouthpieces of the left, destroying their wider credibility, and that's part of the reason we're seeing such a huge right-wing reaction now.
However, we're seeing things like the Bureau of Labor Statistics stopping collecting data, the CDC coming out against vaccines, and giant tariffs and actions that enrage longtime allies like Europe and Canada. This is doing serious damage to state capacity, and as a 'liberal nationalist', to quote Eric Kauffman, this is bad for the country long-term. Overall increases in the level of corruption will take a long time to undo, and I don't think we'll ever get as close to Canada as we were pre-Trump.
I definitely agree they crossed the line going after him over Stormy Daniels. I don't really care about that. However, he did try to launch a coup on January 6, and I wish they'd stuck to going after him on that.
I think the USA should stay a democracy. We weren't perfect, but we were reasonably close to the democratic end of the spectrum throughout our history, and we're likely to get incompetent kleptocrats if we go autocratic. The Chinese have a long tradition of competent bureaucracy checking and moderating the emperor and of the duties of ministers. It's not perfect either, but there's something recognizably Chinese after 2000 years and 7+ dynasties. I doubt we will be so lucky. As for the Russians, well, they've long recognized whatever their nation's strengths it's always been poorly run.
EA assumes all lives are of equal worth, so the damage caused by the end of foreign aid is more than enough from the EA point of view I think. (Personally I thought it was useful as an instrument of soft power.)
there is one point i argue about in context of Israel and now i extrapolate to USA:
There are two ways to understand the word "democracy". The Rule Of The People, and Balances And Brakes.
i think about it as Democracy and Liberalism but i don't think it's proper translation to English.
there are people who see Democracy as the system of governance when people decide by voting, when people are the sovereign, and restriction on it is Undemocratic.
and there are people who see Democracy as system of governance that strongly restricting what the people can decide to do, when the restrictions are the heart of the system. you surely see a lot of this desire to restrict in USA governance!
for people in the second group, the fact that the president is not above the law is important part of democracy. rule that say, for example, execute all Jews, is Undemocratic, even if 80% of the people vote for it.
for the first grope, the decision of 80% of the people is Democracy, and preventing it is Undemocratic.
and instead of talking about this important disagreement, people don't notice it, because those two different principles referred by the same word, Democracy. or even worst, notice and then try to win the semantic game, as if it will let them win the actual game.
(I first read about that in this blog, that is in Hebrew: https://dannyorbach.com/tag/%D7%91%D7%96%D7%9B%D7%95%D7%AA-%D7%94%D7%90%D7%A8%D7%99%D7%A1%D7%98%D7%95%D7%A7%D7%A8%D7%98%D7%99%D7%94/ )
I think you are being too charitable. I think it is true that there are sometimes conflicts between democracy and other values. Some limitations on democracy are good. That is the point of our bill of rights and our equal protection clause. Killing all the Jews would surely be bad, even if democracy were to demand it.
Where I think we disagree is that I don't see the ambiguity in the English language that you do. "Democracy", as I see it, unambiguously refers to rule by the people. Before Trump, and to a large extent even just before Trump 2, that was how it was used. We are now in a world where a lot of liberals, unable to accept that Trump won a legitimate democratic election, twice, are now using the word "democracy" in a dishonest way - they have started using it to refer to "balances and breaks" as you call it, to restrictions. They are trying to challenge democracy because they don't like the leader democracy gave us, and they aren't willing to admit that fact. That is what is going on here. And I will not play into it.
Your problem is that you have bought into the simplistic argument used by dictators the world over: "I won an election, therefore the democratic thing is for me to be able to do whatever I want." This is technically true for a certain definition of democracy, but tends to result in cataclysmically bad government. This also tends to be undemocratic in the sense that it results in outcomes disfavored by most people, like massive economic damage and restriction of freedom. The idea that democracy must be tempered by checks and balances is not new, indeed you may find the phrase used in the Federalist Paper titled “The Structure of the Government Must Furnish the Proper Checks and Balances Between the Different Departments.”
To pick one example of the principle, consider racial segregation. In your sense, Jim Crow laws were democratic in the sense that a majority of voters had voted for them at one point. In another sense, they were undemocratic because by suppressing the vote of a significant portion of the population they damaged the ability of the electoral system to accurately judge what the majority of people wanted. This is why we have a Constitution in the first place: so that one 51% election victory cannot overturn the foundations of society.
The American left is quite aware that Trump won the popular vote (this time at least). The most popular topic of inter-left discourse right now is "Why does everyone hate us more than the most contemptible man alive, is it the fault of the moderates or the progressives?" which very much accepts the turn of public opinion in its premise. The main linguistic compression here is that "democracy" is typically used to mean "liberal democracy", the version of democracy that's patched to avoid demagogue strongmen ruining everything, because that's the version this country has run on since its foundation. More detailed coverage of this topic can be found here: astralcodexten.com/p/defining-defending-democracy-contra
Don't put words in my mouth. I never said democracy was the only value or that Trump should be able to do anything he wants. I took issue with much more specific points that Ozy made. If you disagree with any of the points I made, feel free to respond to them.
I, personally, find Semantic Arguments as confusion at best, and attempts to Sneak In Connotations at worst. I even less equipped to talk about semantics in English, my third language.
but I have one friend that honestly see Checks And Balances as important part of the definition of Democracy, and i don't find it surprising, as i, too, learned that in school. you probably learned something else.
generally speaking, it's the Progressives who choose Sneaking In Connotation as major strategy, and if you will claim that people who believe that is instance of movement believing its own propaganda, you probably right.
and yet, the conclusion remain the same - you just waste time, arguing about semantics instead of content. instead of arguing what Democracy is, you can say that the people are the sovereign, and that the majority should rule, and not some aristocratic minority.
i'm all in favor of naming the thing, the more the better. what can be destroyed by naming it should. but the way to do it is by pointing the substance, not arguing semantics. there are already too much of the debate in the third level of simulacra.
I can't make US political donations, but consider adding Manny Rutinel, running for Colorado's 8th district, to the list? I once had a conversation with him about farmed animal welfare, he helped pass State AI law (not sure how good it was, but an indication that he cares https://x.com/MannyRutinel/status/1798086169290498395?s=20 ) and he's running in a purple district.
Very thorough and well-considered.
This is one of those times where I have to decide whether I'm an ex-liberal angry at feminism, or an ex-conservative angry at corruption.
Honestly, I'll probably do nothing because I'm afraid of retaliation by the administration and I'm a coward.
I want to say I really admire your focus on pragmatism over ideological purity. Yes, it is a good thing to work in the current administration if it means you can push back on antivax!
About the AI... this is one of those things where I don't understand the tech enough to have an opinion, and everyone who does either seems to have a political ax to grind or investments in the companies.
Let me see if I can make a mostly apolitical, non-technical argument in favor of being careful about AI.
PART 1 (assumption). We might be close to actually building something which makes humans economically and evolutionarily obsolete.
This is difficult to visualize or believe, because it has never happened before. But imagine if we *did* build something that could do any task that a Stanford graduate could do for $1/hour: Software engineering, hardware engineering, basic science, being a (good) CEO, operating a robot, designing a better robot, winning an election, lobbying for legislation, consolidating power. Basically, any human task, done well, for $1/hour. Really think about what this deeply strange world would look like.
(If you like, imagine the AIs themselves making decisions, or imagine small numbers of powerful human beings controlling the AIs. It makes less difference than you'd think.)
The AI safety argument rests on the idea that we are potentially "0-2 major breakthroughs" away from this, plus some engineering work. ("Major breakthrough" here means "a paper like 'Attention is All You Need'", which typically come along every 5-10 years.) Sometime between, say, 2028 and 2050, human labor would become absolutely, completely, permanently worthless. Humans would be as economically competitive as chimpanzees.
Will are the odds this will actually happen? Experts disagree. But the AI safety argument is based on the assumption that there's a significant chance this might be a real thing that actually happens within our lifetimes. If you don't buy this idea, then nothing about AI safety will actually make sense.
PART 2 (argument). If a powerful entity and a powerless entity want the same limited resources, the powerless entity is likely to have a bad day.
Evidence: History, economics, colonialism, war, politics, etc.
If it's possible to get any kind of skilled or unskilled labor for $1/hour, then ordinary humans are screwed. Maybe the AIs have their own goals. Maybe the AIs all obediently listen to Billionaire X or Politician Y. But no matter who is in charge, every ordinary human consuming resources is "wasting" resources that the AIs could use better to accomplish the goals of whoever is in charge. The ordinary humans have *nothing* to offer. And if the ordinary humans revolt, the AIs are *also* better at war.
Even if this situation works out for 5 years or 20 years, it's not stable.
So that's my argument: If you make the human mind economically obsolete in every important niche, then it's likely very hard to stop economics and evolution from doing their thing.
But if you don't accept PART 1 (human obsolescence) as a potential risk, then PART 2 (obsolete things tend to get brutally marginalized) doesn't follow.
(There are other possible takes, including "humans suck, let's take a chance on the AIs that replace us." I don't plan to address these right now.)
That's quite reasonable and I've often thought the same. I just feel like a lot hinges on Part 1, and we generally haven't seen the computers be quite as good at blue-collar or pink-collar (service) labor as they are at white-collar lately. Frankly I'm quite pessimistic about the future--I think the billionaires own everything and will just let us all starve.
And on rationalist or rationalist-adjacent blogs I generally assume almost everyone is better informed about this AI stuff than me, so I stick to other topics.
That's exactly how I feel about AI too
Also donate to Mary Peltola in Alaska or the rob sand for IA governor
Hello, Metaculus person here! What would make the US Democracy Threat Index easier to interpret? I'm eager to make indexes more useful.
I don't know how to convert it to a percentage chance that the U.S. is no longer democratic at a particular time!
From the link: "Alex Bores worked at Palantir for a few years. I think that many of Palantir’s activities are pretty evil. That said, as far as I know, Bores did not work on any such projects while at Palantir, and he says that he left Palantir over concerns about its collaboration with ICE."
Working for Thiel for five years, until 20-goddamn-19, and then leaving because of some very vague discomfort over "collaboration with ICE," and *then* immediately running for office as a Democrat raises some obvious questions, such as: "Are you stupid?" and "Do you think I am stupid?" This is a gigantic scarlet letter and I would need to hear an answer, from him, that is specific and convincing. So would the average Manhattan primary voter.
I'm sure he speaks fluent quokkese, but in a seat as safely blue as this one I see little reason to risk a Tricia Cotham situation.
There are many things wrong with the political assumptions in this post, but thankfully the nonsense is clear in two consecutive paragraphs:
"Obviously, you shouldn’t take a position that makes the Trump administration more effective at overthrowing American democracy. But the first Trump administration was an ordinary Republican administration because a huge number of policymakers, advisors, and civil servants thwarted Trump’s authoritarian instincts"
You support unelected bureaucrats having power to maintain the policies they want regardless of the results of elections. But surely you support it because they would stand up for freeedom?
"keeping vaccines somewhat more accessible, making AI regulation somewhat saner, preserving whatever scraps of foreign aid they can"
So this is what you think oposing "authoritarianism" means : ensuring state intervention and regulation.
Your piece barely mentions the policies where the republicans do want intervention and regulation. You want progressives to maintain state control through unelected bureaucrats, and do so in the name of "liberal democracy"
The bit where people are opposing authoritarianism is the bit you left out! But foreign aid, vaccines, and AI regulation are also important issues.
I am baffled by the claim that it is state intervention and regulation to think the federal government should continue to approve vaccines. If we abolished the FDA, any company could put out whatever vaccines they want.
I didn't mention other stuff because I don't give a shit about it compared to democracy, USAID, and AI. For example, Republicans are anti-trans, but as a trans person if polling suggested that a candidate punching a trans person in the face would flip a House seat blue, I would be the first in line to get my teeth knocked out.
New EA aligned research org powerfordemocracies.org has a cost effectiveness breakdown based on ITN and recommends the non profit Freedom2Vote with $94 to increase a voter. IMHO that's more aligned to how EAs would think about impact on the margin.
Generally agree with this, and those of us who work in this space professionally (and think in this way) generally kind of maintain a "supply curve" of different fundable opportunities at different prices (which, of course, can change *within* each opportunity at different budgets). However, there is a massive and important universe of things that can't be measured - things that potentially are more effective than single tactics like Freedom2Vote which can be measured - and so we also shouldn't fall into the "looking under the lamppost because that's where the light is" problem.
Just as an obvious example, a large funder could be thinking right now about ways to increase the chances that the ‘28 Dem primary nominates the best and most electable candidate. You can’t run an RCT on that, but it plausibly dwarfs Freedom2Vote (and I’m a huge fan of Freedom2Vote!).
A better way to convince EA minded donors will be a prioritization framework comparing these candidates against others as well as Freedom2Vote with the funding in margin. I'm sure the industry has some of these supply curve data and analyses and if they're willing to share that could potentially convince more EA minded donors. But in general if you claim of the potential to have a bigger impact without a quant analysis of how much that could be, it's harder to convince EAs.
Orange man is obviously bad (anybody saying otherwise is a psycho or a moron), but he didn't get into the White House twice randomly. Both the Dem and Rep establishments (and their analogues in the rest of the west) are rotten to the core, in thrall to absurd bankrupt ideologies and mainly concerned with zero-sum status games. This state of affairs is unsustainable and deteriorating, and the likes of Trump getting elected is just a symptom. Getting a couple marginally more sane (even granting that they are indeed such) people in office won't change the overall trajectory of the civilization (at best slightly delaying the collapse), and EAs would do well not to fool themselves otherwise.
if you believe that AI may soon come, delaying civilization collapse is pretty important, actually. and even if not, delaying civilization collapse is much better then... what is the suggested alternative? do nothing and watch Rome burning?
>what is the suggested alternative? do nothing and watch Rome burning?
Well, the EA calling is to look for "important, tractable, neglected" interventions, and who's to say that none remain that would be enough to avert the collapse, instead of slightly delaying it?
I like this post, especially the sober Trump evaluation.
I do feel there's some validity to the idea that "There should be an AI risk evaluation committee" really means "I want to be paid to be on a committee and doomsay about AI." But it's a toss-up between who should be trusted less, them or the Tech-Billionaires.