if we're not counting the effort of going to vote, I don't get yud recommending abstaining or cancelling out votes as opposed to voting for trump. it's votes that you add together, so voting trump seems like doing the same thing but harder. the isThreat(x) function is sus here
I asked Eliezer this same question when I talked to him. He said:
"I agree that these two impacts are similar. Similarly, if you're negotiating for an artwork with somebody who values their life at a billion dollars, the impact of "I'm gonna refuse to make this trade which has a gain-from-trade to you of $1M with 10% probability" and "I'mma shoot you with 0.01% probability" are the same to them. But the decision theory surrounding it is not the same."
Unironically, it's explained very well in planecrash if you don't mind reading millions of words of BDSM-themed Pathfinder roleplay with in-world maths lectures.
I don't think we put out any explicit paper on the Nature of Extortion because of the number of unsolved problems in it, combined with a broken journal system that wants a pretense of full solutions.
For some simple ideas going past pure intuition of why "the decision theory is not the same", consider such differentiating observations as:
- In positive-sum bargaining, the other party would rather the whole interaction occur because they have a chance to profit; in extortion, they'd rather the whole interaction not occur. So any sort of coordination work, overhead, or decision-theoretic posture that helps the interaction go through, the other party is motivated to do for negotiations but not for threats.
- A rational agent should always do at least as well for itself as a rock, unless it's up against some other agent that specifically wants to punish particular decision algorithms and will pay costs itself to do that; just doing what a rock does isn't very expensive or complicated, so a rational agent which isn't doing better than a rock should just behave like a rock instead. An agent benefits from building into itself a capacity to respond to positive-sum trade offers; it doesn't benefit from building into itself a capacity to respond to threats.
- Consider the Nuclear Prisoner's Dilemma, in which as well as Cooperate and Defect there's a third option called Nuke, which if either player presses it causes both players to get (-100, -100). Suppose that both players are programs each allowed to look at each other's source code (a la our paper "Robust Cooperation in the Prisoner's Dilemma"), or political players with track records of doing what they say. If you're up against a naive counterparty, you can threaten to press Nuke unless the opponent presses Cooperate (in which case you press Defect). But you'd have no reason to ever press Nuke if you were facing a rock; the only reason you'd ever set up a strategy of conditionally pressing Nuke is because of a prediction about how your opponent would respond in a complicated way to that strategy by their pressing Cooperate (even though you would then press Defect, and they'd know that). So a rational agent does not want to build into itself the capacity to respond to threats of Nuke by choosing Cooperate (against Defect); it would rather be a rock. It does want to build into itself a capacity to move from Defect-Defect to Cooperate-Cooperate, if both programs know the other's code, or two entities with track records can negotiate.
- If Harris had no more ability to respond to threats than a rock, you might offer her your vote conditional on her making a sufficient offer to you, but you wouldn't bother to threaten her with setting off a nuke / voting for Trump (hurting her at cost to yourself), because she wouldn't have built the capacity to respond into herself. An ideal rational agent doesn't want to build a capability to respond to threats to vote for Trump, because it doesn't want to be part of that kind of interaction in the first place; it would rather be a rock in this regard. It does benefit from building into itself a capability to respond to positive-sum trade offers, like offers to vote for it in exchange for its pro-Muslim policy.
As always, do not misunderstand LDT/FDT as making any sort of "commitments" or "precommitments"; the whole point and idea of LDT/FDT is to compute the correct course of action in one sweep via evaluating expected utility using more sensible counterfactuals. It is only other decision theories and viewpoints that need to imagine talk of "commitments" (the places where that other decision theory would rather behave differently than it usually would) in order to see what LDT computes naturally and in one shot. LDT doesn't need to "commit" to paying off the driver in Parfit's Hitchhiker; it just uses more sensible counterfactuals to compute the superior result of paying off the driver.
Thank you! It sounds like you're saying the asymmetry lies in that threats would only be made in the first place if the perpetrator expects the victim's decision procedure to have some pathological flaw that would dispose them to give into such threats, whereas trades would be offered even if the counterparty has a perfect decision procedure.
I am not sure that Cooperate : Defect : Nuke is entirely isomorphic to Vote Harris : Abstain : Vote Trump.
In my opinion, the Muslim's dilemma assumes that absent any coordination, it would be in the interests of the Muslim block to spend the costs to vote for Harris because the benefits of a pro-Israel Harris presidency over a Trump presidency.
If that is the case, then them voting for Harris out of self-interest is the baseline for non-coordinating actors following their self-interest, and any other choice by them would be akin to the nuke option (hurting themselves to also hurt the other player in an effort to gain better outcomes in counterfactual situations), aka as a threat.
The only difference is that there is some cost to the act of voting itself, which is presumably small.
In general, I am still confused about what constitutes a threat that rational actors should ignore and what constitutes a reasonable response to encourage better outcomes in counterfactual situation.
For example, not taking unfair splits of 10$ with some probability seems to be equivalent to the nuke option: you burn some of your utility in order to frustrate the utility function of the other player.
From the payoff matrix alone, you can not determine that a certain split, like 5$ : 5$, is objectively fair. Perhaps both parties have to put an equal amount of work into it and are equally likely to find themselves on either side of the table. But I can also make up a story with equivalent payoffs where one party is an honest trader with a cartload of silks and the other one is a robber (or medieval customs official) who threatens to burn the silks unless he gets a cut.
And in the real world, things are even messier. Should we assume that the utility of 5$ is the same for everyone, or assume that it just increases the net worth of the recipient, and is subject to some concave (possibly logarithmic) utility function? How does the good-will of the pro-Israel lobby (which Harris presumably wants) compare to the desire of Muslims to minimize the deaths of Palestinians? Neither one can easily be translated to dollars.
Maybe tangiental, but those aren't quite equivalent for two reasons. First is almost patronizing to point out, but money doesn't have linear utility, so its exchange ratio with other goods can vary with exchange size. It's under-defined what "valuing your life at $X" actually means, but it'd be reasonable to take the deal "Y% chance of dying for Y% of $X" for some values of Y and not others.
But also here's a broader point about "value of life." I think in this community it's often viewed as "amount you'd be willing to exchange your life for," which implies you put value on money even after death (maybe because you wish to donate it, or have family, etc). But as I understand, it's computed by looking at how much (usually small) risk people are willing to take in exchange for some amount of money. In that case, the point of the money is to use it in the many worlds which you're left living. So those numbers really only make sense in the low-risk regime. Even if utility was linear with money, it would be consistent to be play Russian roulette for $10M, but not kill yourself for $1B, because 5/6 of the time you get to enjoy your money in the first example versus 0 in the second.
Maybe a correction for point #2 is that people have two utility curves for money, one if they expect to be alive to spend it and one if they don't. That I think knowing both would be enough to figure out what risk deals a person would be willing to take.
It's doing something with the same basic effect but different decision-theoretic implications. It's like how if the Vikings demand £1m Danegeld, and your other option is spend £1m to repel them by force, the basic costs to you are identical but the decision theory is very different (you shouldn't give people an incentive to threaten you.)
I'm similarly frustrated by how many states are captured by one party. I also want to vote on a mix of policy preferences, governing competence, leadership, charisma, and integrity. Berkson's paradox grants us at most one of these.
I can imagine systems that might help but would require unpopular constitutional overhauls.
I have decided to vote against the dominant party in a solid state with a clean conscience. I expect that the dominant party should recognize what's happening well before I actually tip an election. You're more likely to cast the vote that makes them think "ooh, that is getting close" than to tip the election. If there are 70/30 splits, you're not voting for the other party, you're voting for more meaningful elections.
This changes by some escalating amount as things get closer. We all kind of know this intuitively, it's why vote trading appears positive sum, votes matter more (read: at all) in close races. If your race is not close, invest in it becoming so. Or, yes, apply a declining probability of doing so as it approaches 50/50.
Actually if you're serious about this don't just vote contra normative, put up a billboard urging others to do so as well. A single vote resembles buying a lottery ticket and pledging the winnings to your chosen party. ("Don't defect," they say in lottery elections world, "it only works if everybody does it...")
Unless you believe the majority "bad governance good policies" party in your particular state will take the wrong lessons, adopting bad policies while retaining bad governance... that is a risk!
There's also an empirical question about whether parties even act on defecting constituents at all, or whether they are goldfish brained random walks. Immediately after 2000 did third parties become more politically powerful or more ostracized? I think further ostracized... rough news for their nash equilibrium.
If you actually want to impact party decision makers, you probably need to move up the chain and do it through media or money. The mechanisms through which Biden was replaced provide an instructive case study. It involved a somewhat quiet and conditional donor revolt backed by arguments that this was actually the right call, with levers of gradual escalation by making it more public over time.
Can you spark a donor revolt over any of the things you find distasteful? How would you position yourself to do so? (Once you are positioned to do so you will probably find the weapon cannot fire repeatedly and you are stuck in a future resembling the one you were already headed toward without all that effort..)
Arguably though the initial seed on replacement was planted by Nate Silver, so punditry is possibly the most sustainable a vehicle here, even though a longshot much of the time.
Dunno about Eliezer, but to answer why voting 50/50 and voting Trump are different: US presidents are still elected by a minority of the country. Elections aren't won by convincing the other side you're worth voting for; they're won by convincing your own group to actually get off their asses and cast a ballot. If 1,000 people vote 50/50, that doesn't sway the election in the direction of the opponent, but it demonstrates that by acceding, you can get 500 fewer people voting for the opponent *and* 500 more people voting for you, who would otherwise be staying home because they don't want to vote for someone who goes against their core values. And since it doesn't *actually* sway the election in favor of the opponent (i.e., it doesn't on net directly go against their interest), it's much easier to convince your otherwise-preferred-candidate that it's a threat your group will be willing to follow through on than having to convince those same 1,000 or even 500 people to vote for the opponent without a counter-balancing trade.
I feel like framing them as 'defecting' is extremely unfair. They were offered an unacceptable deal from their point of view. Despite being a reliable voting block in the past their concerns were not given much weight.
People being willing to 'swap sides' provides good incentives to those in power. Even if the new guys are worse, if you totally screw us over we might side with them...
If we frame the election as a process that collects information about leader preferences and then selects the best leader on the basis of said information, then we might think there is a obligation under the social contract to feed accurate (to the best of one's knowledge) information to the system. If so, then the aforementioned Muslim protest voters are defecting not from Harris, but from democracy; they are behaving unethically because they're engaging in deception.
(compare being paid to vote for someone, or being paid to write a glowing letter of recommendation for a terrible employee applying to a socially important position, or writing said letter just for the lulz)
SHOULD we frame it that way? I don't know, but my impression is that political scientists tend to see protest voting as caused by inadequacies in the information collection system (e.g. maybe voters are forced to vote for the lesser evil under FPTP but they would be able to vote their conscience under RCV). So here protest voters aren't deceiving anyone; they're communicating imperfectly accurate information to a system that doesn't let them articulate it better. I don't know which inadequacies of this kind would cause the Muslim Trump protest vote; if there aren't any then this is a substantially different case.
Not sure if it makes sense to bring up ethics under a game theory post but maybe this is a good way to understand why someone would disapprove of their choice despite understanding their concerns.
This feels pretty gross and off-topic, and also, who could possibly answer it? I lived in Michigan most of my life and not one of the Muslims I knew was in favor of terrorism. Is that useful information to update your 'curiosity'?
Why is it gross? It could be answered with 'Based off knowing people in Michigan, 99% would condemn this attack. Lizardman factor always exists but 99% of them are very decent people' [hypothetically!]
I think it's not a fair question because nobody could possibly claim to know a representative sample of Muslims in Michigan with any real authority, and I think it's gross because it implies it's totally reasonable to assume there might be a ton of people living in Dearborn who are in favor of terrorism.
And, as I said, it's not really on-topic for this post - the post is about whether it's reasonable for people who are concerned about the fate of Gaza residents to effectively support Trump to increase their political leverage, and your response to that was to ask how many of them are pro-terrorism.
All discussions here have spin-off discussions that are pretty far from the original Scott post. And I think it's OK for people to ask gross questions here, questions that imply that a certain group of people is dumb, genetically violent, whatever. Or questions that come right out and say this or that group is savage and inferior. There are enough people committed to "where's the evidence?" that there's no need to feel concerned that some somebody's implied negative view is going to plan the seeds of irrational prejudice in a bunch of readers.
Eh, maybe. My feeling was it was a low-quality comment in terms of how inflammatory it was vs. how effectively-impossible it would be for anybody to give a genuinely meaningful answer to it. I'm not asserting it violates any rules, I'm saying I expect it to drag down the discourse into a bunch of competing anecdotes that make people angry without enlightening anything.
Been lots of people on open threads making the case that blacks are dumb and savage. And people routinely say devaluing things about women here. I put up a post with multiple examples of the latter recently. Latest instance that caught my eye was somebody's sketch of how we might use embryo selection and breeding to improve human stock: Brilliant men marry beautiful women who go along to get along, and also put out -- no, wait, misquoted the last part -- I believe the exact words were something like "have pleasant temperaments and are mentally stable."
My question is reasonable because it’s relevant to whether this group is fighting for Gaza from a place of compassion and kindness or from a place of hatred towards the other guys. This makes a big difference in how sympathetic we should be to their political demands.
My model is that most groups focus on the atrocities committed by their out-group and tend to ignore the atrocities of their in-group or far-group.
Marxists might talk at length about the genocides of the fascists and not a lot about Stalin.
Republicans talk a lot about violent immigrants and not a lot about school shootings.
Democrats talk a lot about police violence and not a lot about murders committed during the BLM riots.
The default position to take with regard to atrocities which don't fit into your groups narrative is to ignore them, sprinkled with 'thoughts and prayers' and 'let us not politicize this tragedy' to taste.
It is hard to say to what degree people intrinsically value the lives of their political or ethnic opponents, because indirect ramifications dominate. For example, I don't like Trump. However, I also think that political assassination is terrible, so I am glad that the shooter missed. Then again, if he dropped dead out of natural causes tomorrow, I would not mourn him, because I think his political effect is net negative and much larger than the value of one life. However, if I could either cause him to cease existing or being isekaied to a simulation of the White House (which is otherwise devoid of qualia) where he gets to play president, I would rather send him to that later place than condemn him to non-existence or hell.
I have one Muslim friend in Ontario who shared a story that seemed to say that the Jewish provocations before hand, e.g tearing down a Palestinian flag and hitting a taxi with a crowbar, were valid justifications to hospitalize five people.
Unfortunately most people do when they're angry. Or maybe "badly-target punishment" where you punish innocent members of a group for offenses other committed.
I'm Muslim, albeit not from Michigan. My honest answer:
Probably a higher percentage than of the Jews in that crowd who would even implicitly condemn the well-documented war crimes of the IDF. In fact, they _explicitly anti_-condemned such war crimes by chanting, “Why is school out in Gaza? There are no children left there.”, literally cheering on the deaths of their opponents' babies which their kinsmen are perpetrating.
By the way, those are called fighting words, and they're not protected anywhere AFAIK. Imagine Muslims anywhere cheering on 9/11, or worse, gloating in ISIS murdering the babies of the kuffar. Would anyone be asking for the condemnation of those who attacked that crowd, even if we charitably stipulate that the questioner assumes in their question's framing without substantiating it that innocent Muslims were dragged into the violence?
Anyway, I also have a hunch it'd probably be a lower percentage than Muslims in Michigan who'd implicitly condemn any attacks on innocent Jews who didn't chant genocidal taunts, but a higher percentage than of Jews in Israel who would explicitly condemn the well-documented war crimes of the IDF.
Now I have a few questions for you. How is this on-topic? Do you find yourself questioning the condemnations of any other groups, and if so , which?
Can you explain how your on the surface unkind (accusatory), not true (disingenuous), and unnecessary (off-topic) question is actually any of those things? — and if you start by saying that it doesn't matter that it's off-topic, maybe you can also tell us about how you wanna ask the same questions of other groups in other threads, or explain how it's so necessary to ask these off-topic questions of Muslims but not other groups.
Are you fishing for condemnation? Why don't you lay out the pertinent facts as you see them when asking people about what they or their community would or wouldn't condemn, to make clear what you're expecting(?) or fishing for(?) them to condemn (no, linking to a Wikipedia page doesn't count, even if it weren't a live page)?
Finally, on the meaning of “explicit”: I've been operating under the assumption that by “explicit” condemnation you meant going out of one's way to make a statement that one condemns something before being asked, as opposed to simply being against it ideologically and/or emotionally but refusing to play the condemnation game, for example; is this what you meant?
I've also assumed that the middle point between those two positions — whereby one makes one's opposition to something clear when asked, but doesn't go out of their way to make a statement beforehand — as being implicit condemnation, too; is this what you intended?
Explicit = if asked, they'll say "That's highly disappointing" or "I condemn that" or "I hope the authorities justly prosecute those involved".
The opinions of American Jews on Gaza are not something I'm interested in, simply because I know a lot of them personally and don't need the Internet to provide me with an answer.
You've written a very long explanation of the relative degrees between Muslim condemntation of the attacks in the Netherlands and the Jewish condemnation of the attacks in Gaza. But you never once listed any numbers. What is the exact number (with error bars, if you must) of Muslims in Michigan whom if told about the events in the Netherlands will explicitly condemn them, even in a private circle of their friends? It could be 99.9%. Could be 50%. Could be 5%. I have no idea, so you tell me.
You (and others) in this thread seem very uncomfortable with the question for mysterious reasons. The default assumption is that ~99% of Muslims will condemn the attacks in the Netherlands, correct? If so, why be nervous, just say "yeah, 99% will condemn" and that's it, end of story. Ask yourself what makes this question so, so inappropriate to you.
>The opinions of American Jews on Gaza are not something I'm interested in, simply because I know a lot of them personally and don't need the Internet to provide me with an answer.
I think necessary is supposed to be about important questions that need to be asked, not questions that _you_ personally don't know the answer to, or requests for confirmation of a community's decency who you have little faith in.
Why did you even think the comments of this article were the place to get your answer? It’s a pretty specific question about one religious community’s reaction in one American state to a very recent event, you’d probably have to actually do the polling yourself.
>You've written a very long explanation of the relative degrees between Muslim condemntation of the attacks in the Netherlands and the Jewish condemnation of the attacks in Gaza. But you never once listed any numbers.
Even those relative degrees, I hedged with probably. As I stated, my honest answer. They’re my guesses based on my experience of Muslims I know and seeing things like the mockery of dead Palestinian children by Jews becoming normalised, even if not accepted.
I couldn't imagine such a large crowd of average Muslims shouting the things those Jews shouted, they'd have to be an ISIS brigade or something.
But I don’t know the Muslims in Michigan. I haven’t asked them. I don’t even know if they’d agree with you about what the facts of the event were.
The reporting on that event was enough to completely discredit the mainstream media to a community that already distrusts it when it comes to reporting on their community, especially in the current context of the ongoing genocide and the increased need for propaganda against them.
Simply put: they may not explicitly condemn it even if asked simply because they’re sick and tired of being constantly treated by the media in a way where their only two options are a) validate the media narrative, or b) look like villains.
They don’t care about the latter anymore, especially with public support being with them as the public see through the media distortions and have come to simply immediately distrust any reporting with Muslims involved where they’re painted as total villains while their ‘victims’ are completely innocent.
Of course, many journalists across various outlets have spoken about the top-down restrictions on using terms like “Palestinian”, being forced to use alternative terms like Arab instead, whereas they're allowed to say “Israelis” or “Jews” as they please, they've been exposed with the constant double standards in the use of active vs. passive voice, or infantilising teenage female soldiers of the IDF while adultifying literal toddlers (my favourite are the IDF bullets which “found their way” into a car and killed a “young lady” of toddler age), etc.
But Muslims know from daily experience how far removed the stories in the MSM are from any Muslims around them. They simply wouldn't believe the allegations of a so-called pogrom of Jews because it goes so counter to their experience of actual everyday Muslims.
So I want to stress again, a good chunk would simply not condemn it because they just have such little belief in MSM narratives when it comes to their community. And there is a widespread belief among Muslims, and I think now among non-Muslims that the MSM has a strong bias in favour of Jews and against Muslims, especially on certain topics.
They may seem to have a bias in favour of Muslims at times, but that's a bias in favour of liberals who are Muslims in name only, Muslim as an ethnicity, not as a religion.
Anyway, the point is, we know they're liars. It's no different to me believing a known and compulsive liar that I know personally and who has been exposed numerous times in brazen lies. Why should it be any different if he's got some markers of legitimacy like a PhD or a Nobel prize or whatever else, when I know him intimately and know what a fraud he is? Why look at the proxies to determine the reality when I know the reality?
I honestly genuinely couldn't give you great error bars. Higher than 5%, lower than 95%, with >75% confidence. There are just so many factors that go into condemnation beyond the content of one's beliefs.
If you want the actual belief content, every Muslim I know would be deeply opposed to attacking Jews simply for being Jews. Reports of Muhammad genociding Jews are greatly distorted by people like Douglas Murray.
I don't say this as a liberal who's Muslim in name only. I believe we should lash fornicators, stone adulterers to death, chop the hands of thieves, I'm a Muslim, and I don't pick and choose.
>You (and others) in this thread seem very uncomfortable with the question for mysterious reasons.
No one's uncomfortable, they're disgusted with you. If it's a mystery to you why that's the case.
>The default assumption is that ~99% of Muslims will condemn the attacks in the Netherlands, correct?
If that were your default assumption, why even ask? The default state of interacting with humans is that we assume the best of them until proven otherwise, without going to the extreme of unnecessarily trusting them with everything of value.
To start questioning the integrity or decency of an individual or community requires some cause for suspicion, so naturally, when you do start questioning people's character, there is an implied attack on their character.
There's no excuse for this level of cluelessness. I think you know all this.
>If so, why be nervous, just say "yeah, 99% will condemn" and that's it, end of story. Ask yourself what makes this question so, so inappropriate to you.
No one's nervous, we're not gonna give you an answer out of thin air to a very specific question regarding a recent event where the media lied (and are known to lie in general), we're not gonna validate your hidden premises (e.g. “will you condemn these attacks on Jews” obviously has an implied “innocent” there), we're not gonna accept your hiding of those premises and framing of the question in such a way that gives only the option of condemn and validate a false implied narrative or don't condemn and validate a false image of us as desiring to hurt Jews simply for being Jews or not caring about injustices perpetrated against innocent Jews.
>Ask yourself what makes this question so, so inappropriate to you.
I'm well aware of what makes it inappropriate to me. You haven't caught me out in some kneejerk moment where I have an automatic/unthinking negative emotional reaction to someone without really knowing why.
---
I'm gonna repeat a few questions you didn't answer the first time.
1. How is this on-topic?
2. Are you fishing for condemnation?
3. Why don't you lay out the pertinent facts as you see them when asking people about what they or their community would or wouldn't condemn, to make clear what you're expecting(?) or fishing for(?) them to condemn (no, linking to a Wikipedia page doesn't count, even if it weren't a live page)?
And a few new questions?
4. What is _your_ perception of the opinions of American Jews on Gaza w/ percentages of various camps?
5. What is _your_ perception of the opinions of Israeli Jews on Gaza w/ percentages of various camps?
6. What is _your_ perception of the opinions of American Jews on the spate of Muslims being attacked or killed by Zionists, Jewish and non-Jewish, including the little boy who was killed by his mom's landlord who'd always treated them well before recent events, or the Jewish woman at a swimming pool who drowned a Muslim lady's baby in the pool?
7. Which of the above incidents, or similar, were you aware of?
8. Do you condemn any of the above incidents?
9. What do you know about Islam?
10. Who taught you what you know about Islam?
11. Do you know that God doesn't command genocide in our holy book, so you have nothing to worry about when it comes to Muslims being sociologically far more adherent to their texts? This is unlike some books, such as the foundation of the oh-so-tolerant Judeo-Christian Western civilisation who were “the first” to “abolish” slavery.
1. It's a post about the political opinions of Muslims in Michigan about an ongoing war between Israel and Gaza. ~95% on topic
2. No, you're fishing for one, I'm asking a question.
3. I asked a question, links are the responsibility of those providing an answer
4. Very roughly: 90% would condemn/are upset about the civilian casualties in Gaza, 10% support them.
5. No idea, I don't live in Israel
6. If it happens in Palestine I'd say 90%/10% condemn/approve split would apply too. If the happens in the US or the Netherlands then probably 95% would condemn it, since the Palestinian question has been polarized by 10/7.
7. None
8. Of course
9, 10, 11: Islam cannot be held responsible for the actions of individual humans. People have free willl, they can't hide behind relion.
I really enjoyed this analysis; however, in this specific case I think it is missing the fact that some of these Muslims actively thought Trump was better (he has done effective outreach to them on how he will 'stop the fighting'). Now, I think they are probably wrong about that, just as the people voting for Trump because he will stop inflation are probably wrong, but that doesn't change the fact that - at least some - are voting for a candidate they see as better, not just against one they are angry with.
A journalist friend of mine spent some time with the Arab Americans in Dearborn over the election and wrote this about it:
I came here to say something like this. My thought (just from reading articles and stuff) is that it's not that Arab-Americans and Muslims particularly like the Democrats, but that they really hate Neocons. Dick Cheney endorsed Harris, not Trump. It wouldn't be surprising at all if just that endorsement repelled a significant percentage of Arab-Americans away from the Democrats. You don't need complex game theory: one of their worst enemies endorsed Harris.
> (TLDR: Trump is bad, but Cheney is world-historical, genocidal evil bad)
I find myself roughly agreeing with the AJ article that the Bush era war on terror was a catastrophe on many fronts. The long term outcome of either the Iraqi or the Afghanistan invasion were bad. Gitmo was terrible.
However, when I think world-historical, genocidal evil my yardstick is Hitler, and GWB and his cronies are just not playing in the same league, evil-wise.
Gitmo is evil, but it is not Auschwitz in either scale or intensity. GWB has earned his place on the list of post-WW2 leaders causing large scale human suffering, but without having checked the numbers, my gut feeling is that Vietnam was likely worse.
In particular, I think that the framing of Cheney as genocidal is inflammatory and thus should be backed up by evidence. Even though AJ is not shy about making inflammatory allegations of genocide (e.g. against the IDF), they do not make such allegations against Cheney.
Yes, I think this probably explains more of what happened than some complicated game theory take. I haven't seen a ton of exit polls specifically breaking out Jewish/Muslim voters, but Trump winning Dearborn seems pretty surprising to me. It seems that the Biden/Harris administration got themselves blamed by all sides (perhaps unfairly); but at the same time, it's not surprising that voters would say that Trump's 1st term saw a much more peaceful Middle East than Biden's term, and for them to vote accordingly.
What we didn't see was any sort of decisive break to the Green Party, or any other third party choice. I don't think Trump won any states where a combo of Democratic+Green votes would've beaten him. It seems to me that voting third party would be a better way to signal your power as a voting bloc: if Trump's votes go up, you might just assume he turned out his base better, but if an interest group comes out and says "We're voting for Stein" and then she does way better than the Greens normally do, you can more easily chalk up the difference to that one interest group.
I think this is really it. An anti-incumbency vote against the administration that perceived to be prosecuting the conflicts with a sprinkling of cultural topics contributing. Harris failed to distance herself from Biden so they voted for the other guy.
And religious people tend to be conservative + conservative people tend to vote for Republicans. Before 9/11 and the war on terror Muslims leaned towards Republicans.
I read a twitter take that was something like this:
For years there's been jockeying for status among the Democrats' intersectional coalition. But recently two clear winners have emerged: black women, and transsexuals, with all other groups demoted to a subsidiary role within the coalition. As a result a bunch of other groups (this particular tweet was about Latino men) are abandoning the Democrats.
> And religious people tend to be conservative [...]
Conservative in what sense? The current Pope is rather religious [citation needed], and he sounds like a commie (in line with what most of the Bible says).
I don't particularly get formal decision theory, but similarly feel somewhat flummoxed and disappointed at the perennial non-catered-to-voter's choice of "settle for whatever slice of loaf coalitional politics is willing to toss my way" or Threaten To Vote For Someone Else. It can definitely lead to some deeply uncomfortable places to be in that taken-for-granite bucket, being expected to go along with outrageous nonsense you wholeheartedly disagree with because You Should See The Other Guy / No Candidate Is Perfect (said in patronizing voice by a Very Serious Adult who compromises with sin). And, yeah, in places like CA where the opposition party is mostly a formality, there *isn't* always a Someone Else. Or they're so laughably below A Literal Heuristic Rock that one is inclined to simply not vote (but then that weakens the ostensible feedback mechanism). Organization helps, to some extent, but...that just kicks the coalitional politics can down the road. Building a smaller tent out of those who self-select or are rejected from the bigger one is gonna have some gnarly selection effects by default. No orthodoxy for the heterodox!
(Vote-swapping compacts seem like a clever solution to this dilemma, but pose other large coordination difficulties that I think in practice make them impossible to scale effectively. The mercenary attitude necessary for being willing to bargain one's vote necessarily makes counterparty trust a bit tricky.)
Voting isn't really the main vehicle by which most interest groups get what they want. Small but organized groups can extract benefits for themselves at a cost to the overwhelming majority, usually through means like campaign donations, lobbying, protests, getting their members to call or write to politicians, getting their members appointed to boards and departments, using their expertise to suggest regulations or laws, etc. Also, in many cases they can do these things regardless of which party is in power. The ability to deliver votes is usually minor in comparison.
> So I think this is the right move if everyone involved is a superintelligence and has already made all commitments they would make if they were omniscient and had infinite time to think about the problem.
I don't think it does work, because exactly the same logic applies in reverse to voting for Trump. Trump's offer is even lower, so they should punish him by voting for Kamala with even higher probability than the reverse. And in real life, if you ask the leftists who keep trying this strategy why they don't try to influence Trump instead, they say it's because they think Trump can't be influenced to support them, and the Democrats can, which seriously undermines the idea that this isn't threat-like behaviour.
(The Palestine-supporters are by no means the first leftist group to try this, I've been hearing leftists talk about how they should punish the Democrats for not being leftist enough since at least 2016, and the Palestine-supporters almost certainly got the idea from there)
Maybe that's why Eliezer recommended abstaining or canceling our votes instead of voting for Trump, which maybe wouldn't have this problem? I admit I'm very confused about it though.
I think Eliezer's notion of "threat" relies heavily on having a privileged null action. If you pick "don't vote" as the null action there's no longer a contradiction, you just don't vote for anyone, but I'm not convinced that really makes sense either.
No it was not and it has never explicitly been that until maybe this very election, and just barely. That's what it "should" have been about perhaps, but instead it was mostly a completely moralistic argument that you shouldn't have to vote for someone with whom you disagree on almost everything, combined with some vague hope that a third party could come to power, which very few take seriously. I've never heard a leftist actually articulate an argument to vote the opposite way until just a few weeks ago in this interview with Noura Erkat: https://youtu.be/bqQ75cktTY8
"I want Kamala to lose". Still it was very rare to hear people expressing stuff like this publicly and many refused to vote, voted third party, or voted for her anyway. What's different about this time? Everything. Kamala made 0 concessions to anyone left of center and in particular to those against the Gaza war, none, when asked how her administration would be different from Biden's she answered that she couldn't think of anything.
Really, to a leftist it's 99% to 100% Satan, not just a cute thought experiment, there's a utilitarian argument to be made that settling for that extra 1% is worth it to get someone who's double digit not Satan in 4 years, and to hold them accountable with the threat of voting Satan if they try to sidle up to him again.
I personally saw leftists in 2016 insisting there would be no reason for them to vote if Bernie Sanders wasn't the candidate, or if the Democrats didn't adopt a socialist platform, and making similar demands in 2020. AFAICT nothing has changed here except the specific demands. This is particularly notable because Biden in fact made substantial concessions to the Left by e.g. passing a huge economic stimulus, dedicating a bunch of money to renewable energy, trying to forgive a huge amount of student loans only to be blocked by the (republican) courts, and to my observations this didn't reduce the amount of Leftist demands at all. I even recall people explicitly saying they would respond to policy wins by increasing their demands.
It seems to me that the people making these demands have, at best, drastically overestimated the size of their voting block, and more likely are just more committed to railing against the democrats then they are to accomplishing any policy outcomes, and either way there is no reason to make any concessions to them at all.
Oh they absolutely did talk about not voting. In the end though it was the same argument, they couldn't bring themselves to vote Biden because they recognized he hated their guts and wouldn't do anything substantial despite some token rhetorical concessions. This is actually what happened btw if you were paying attention... And anyway a lot of these people, probably most, would later admit to voting Biden anyway.
Regardless, I'm talking about a strategy of voting the other side, assuming you're in a swing state, as in you are explicitly voting for the outcome you desire which is for the person you hate even more to win. You would never have heard this mentioned in any kind of even remotely public left discourse in 2020, it just barely started happening this election and only right at the very end.
There are always leftists who say "we won't vote for the Dem because X" and X changes from student loans, to universal health care, to Palestine, and even if that was solved there would be a brand new thing.
The centrists have a much better chance of bluffing. They are, by definition, less bothered by a Trump victory. Biden's position was closer to what they wanted than Trump, but even Trump's position isn't unpalatable.
>“here are the usual levers for influencing Democratic Party policy, like voting in the primary, you lost, and you are trying to re-litigate a completed political process because you don’t like the result, in the same way we condemn Donald Trump for doing.”
This argument would have been more effective if there was a competitive primary this year. (for the record, I voted for Dean Phillips, but the best potential Democratic candidates didn't run out of deference to Biden)
I am very confused why this is framed as decision theory rather than game theory. If you view it as a repeated game, standard game theory tools make sense here. In particular, accepting one cent isn't necessarily a Nash equilibrium of the ultimatum game if the game is repeated; you make yourself worse off now to get better offers in the future. Muslims are doing the same thing this election in this example; they're making themselves worse off now in the hopes of getting better offers in future elections. I really don't see why you need to throw on so many epicycles to explain it.
Exactly! I was thinking the same thing. Eliezer's answer seemed overcomplicated to me (also I really dislike it when people like using so many technical terms that it obscures a lot of the thought process). Yes, in artificial experiments where you know the game doesn't get repeated it is rational to accept 0.01. But this almost never happens in reality and even if a particular game doesn't get repeated, you will play other games and as long as your behavior can be observed of any party you might play new games with, it is still better to refuse to maintain a reputation of someone who cannot be mugged in a game like this. And perhaps more importantly, this is mostly instinctual behavior in most people most of the time (and not just people, some smarter animals have this hardwired too) and it is not worth thinking to hard in an experiment where the total prize is negligible. I would expect different results if this is a prize of 1 billion and you're offered 1 million or nothing (also because a game with those stakes is unlikely to repeat and it is so obvious that you act accordingly).
I also agree that Yudkowsky’s answers are low signal to noise ratio and these seems like a pattern with him. Of course this is a repeated game, and defecting makes perfect sense as long as your gripes with the Republicans are not existential.
Actually, as someone who was in that massive Discord thread, I think Eliezer's answers here are still less specified than I want. This stuff is tricky, and there's a lot of moving parts, and I only understood some and have forgotten much of that.
Like, one of the things with your $999m to $1m split is, how aligned are you with the other player? A little quote that I'm proud of: "the greatest trick the devil ever pulled was convincing the world that non-iterated games exist". To a close approximation, you'll always have another interaction, and how confident are you that the other player won't use that $999m to act against your interests in a way that you can't resist with a measly $1m?
> Eliezer's answer seemed overcomplicated to me (also I really dislike it when people like using so many technical terms that it obscures a lot of the thought process).
I suspect that's because he assumed a lot of shared context that you don't have?
No, it’s just really not a good answer. Lots of jargon and epicycles in his examples, but not very illuminating versus simply applying *really* basic game theory. Even Scott doesn’t seem to find it very satisfying (“I think that’s what Eliezer means…”) though I think for personal relationship reasons he extends Yudkowsky a lot of credit that he has not earned in this realm.
I agree that in this case, modeling it as an iterative game is fine, but in general it's still worth understanding why it's good to cooperate with the stranger you meet only once and never again.
I think one of the points of the Yudkowsky style decision theory is to somehow merge decision theory and game theory. If we treat all the different players as somehow algorithmically correlated, we then just have to optimize that algorithm against the background of whatever is left, rather than treating the different players as separate.
I don't think the war in Gaza is a genocide (though it might be some sub-genocidal level of bad, IDK), but if you did think that and you identified with the victims, it would probably be more important to stop/mitigate that in the next four years than get whatever unknown and hypothetical concessions you think you might be able to get after the genocide is completed in the next election.
First of all, yes, in general, it depends on the relative values of what you're losing now and what you might gain in the future. It could be the case that preventing harm now is more important than future gains, but it could also not be. In particular, it could be the case that the belief is Trump will make the war worse, but it's going to be bad either way, and the national impact of Trump is small. In that case, building the reputation and making future gains could easily outweigh the present harm.
>In particular, it could be the case that the belief is Trump will make the war worse, but it's going to be bad either way, and the national impact of Trump is small.
I agree with this, and that this would go a long way to explaining the phenomenon in question. But for the record, it seems like a just false empirical belief to me, and most of the people accusing these voters of being irrational (rightly or wrongly) are trying to point out to them that the war will be much worse for Gazans than it already is under Trump. I guess we'll see.
Yeah, in a sense this is the entire reason behind war? People are willing to potentially sacrifice their entire future to prevent something bad enough from happening right now.
In the repeated ultimatum game, played by two CDT agents, the last round plays like the single-shot version. Since they already know how the last round plays, they know anything that happens in the second last round has no impact on it, so it also plays like the single-shot version. By induction, every round has the offerer offer the minimum value, and the decider accepts it.
This is essentially the same reason why CDT agents mutually defect on every round in the repeated prisoner's dilemma.
To get some other outcome, you need some other theory of decision making, that cares about things other than just the causal effects of its actions.
I hate to spoil the nerd party, but has it occurred to anyone that maybe the Muslims are genuinely just mad and that there isn't a game theoretical reason for defecting?
( anger itself is somewhat game theoretical, but the incentives acting on an individual are about how they think other people will view them based on what they've heard on social media, which is Game Theory but with such skewed and distorted inputs that it's really more of Sociology than strategy.)
Not to mention the fact--easily ignored on a place like this but highly salient--that most Muslim Americans motivated by that identity are varying levels of highly religious, and therefore believe that their deity is directing matters toward an ultimate end whose shape and characteristics they already know (Judgement Day).
On a related note, most religious Muslims are highly conservative, and only support the Democratic coalition because they expect the Republicans to oppose Muslim religious freedom and immigration. In terms of other social issues, they tend to be as opposed to Democratic policies as conservative Republicans are. Democrats are very much an ally of convenience for them.
As you say, evolution obviously programmed us to feel anger for game theoretic reasons. And I think it's worthwhile to understand why evolution did that and maybe spot places where we shouldn't indulge in our anger. Maybe the Michigan Muslims don't care about this question, but *I* care about knowing whether their strategy is correct from a game theoretic perspective.
I follow one of the people pushing this movement, and it’s pretty clear he’s approaching it from a game theoretic perspective, even if not formalising it in such terms to his audience AFAIR.
It’s not just a phenomenon of American Muslims lashing out due to anger, they’re thinking about the issue from a rational leverage perspective.
This is the most rationalist sentence I've seen you post in some time, but I feel like if you don't qualify it it'll definitely show up in the next newspaper profile of you.
Fair point. I would throw a similar idea out there: I work with a Palestinian guy (n=1 but I don't think he's all that unusual). He is, as you might expect, highly socially conservative. You can imagine someone like him being willing to cast his lot with the Democratic Party despite his deeeeeeep disagreements with them on most issues if he feels it might be of some benefit to his family back home. Take that away, though, and he's basically a Moral Majority devotee. American Muslims: not single issue voters!
If you immediately know the candlelight is fire, the meal was cooked a long time ago.
In other words, a solution that avoids this problem is, "Start organizing and advocating in useful directions in 2004-2016, when it was already becoming clear which direction things were moving and that the democrats lacked a deep bench of candidates that were actually able to competently represent their constituents." Then, in the cases where it happens that you still get a candidate you can't stand, you can either support them anyway without such long-term decision theoretic implications, or vote against them and trust that there are enough other strong democratic candidates in other offices that you're not throwing the country to (what you may see as) the wolves.
We look around us right now, at trends in both major parties, and start trying to influence (or create) politicians in both parties who think about things in similar ways.
The Ultimatum Game comparison seems flawed because it's a positive-sum game. Both participants can win money if they find an agreement, or nothing.
The voting situation regarding Israel seems negative sum to me. Harris can either appeal to pro-Palestinian voters, and lose pro-Israel votes, or appeal to pro-Israel voters and lose pro-Palestinian votes. Assuming that there are more pro-Israel than pro-Palestinian voters for Harris to lose, then (a purely cynical version of) Harris isn't gonna do anything for pro-Palestinians no matter how hard they threaten to vote for Trump, and voting for Trump is a pure waste that doesn't give anyone incentive to help them.
I think this is a common failure more of leftist ideology when confronted with centrist blocs. Leftists often show contempt for centrism/pacifism because they assume the only possible option a reasonable person can have is "do nothing" or "side with me", and they resent people for picking option 1.
That doesn't make the situation negative-sum for Harris, just one with tradeoffs. Harris can still set a policy on Israel that she expects to get her as many voters as possible. Voting blocks can try to shift where Harris falls by partially or probabilistically withholding their vote unless Harris crosses some particular threshold.
(If the voters are naive like CDT and think they have to vote for "whichever electable candidate has the better Israel policy", and don't consider "don't vote / vote 3rd-party" as an option, then Harris can collect all of their votes via the tiniest possible policy shift away from Trump.)
It's true that that doesn't necessarily make the situation negative-sum.
But then again, politicians often feel that taking a position is a negative-sum thing to do - that any position they take will get them less support than a positionless feel-good platform - and I suspect they are usually right about that.
There's a scene from the sitcom Friends that's stuck with me:
In this episode, the audience learns that Phoebe is secretly married to a Canadian ice dancer, Duncan, for purposes of making it easier for him to live and work in the United States. The marriage is one of convenience; he proposed to her as a gay man who would have no relationship with her other than the existing friendship and a formal state of marriage.
But, we learn, Phoebe agreed because she was secretly in love with him. The reason we learn this is that Duncan has shown up to ask Phoebe for a divorce - he's realized that he is straight, and he wants to marry a different woman.
Phoebe is naturally upset. And when she agrees to the divorce, she can't stop herself from asking "if you had realized you were straight earlier, do you think I would have been the one..."
At which point she interrupts herself, saying "you know what, I don't think either answer would make me feel better."
This seems to be a plausible example of a situation where knowing more about it always leaves you worse off, regardless of what it is that you've learned.
> The Ultimatum Game comparison seems flawed because it's a positive-sum game. Both participants can win money if they find an agreement, or nothing.
We typical compare fixed sum games and variable sum games.
We typically call the former 'zero sum' games, but any fixed sum yields the same analysis.
For variable sum games, applying any common additive offset to all outcomes doesn't change anything about decisions. (That's like: 'you find a coin on the pavement before the game begins.')
Scott, what’s happened to you? You’re ignoring all the reasons Muslims might actually _like_ Trump and adopting, without any questioning or criticism, the brain dead narratives of the New York Times crew.
The man expanded his coalition among basically all groups, because he actually tried to do this. The Harris campaign’s whole argument was, “orange man bad”, and Trump said he was going to end the conflict quickly. Doesn’t matter if you think the guy doesn’t like your religion - what matters is he’s at least saying the killing has to stop and you suspect he might actually deliver.
Rather than assume I'm trusting any "narrative", you can just listen to what these people say. https://abandonharris.com/faqs/ , second video, "What Happens If Trump Wins".
I understand people are getting overly twitchy because of the election result, but I request an apology for your first paragraph.
So there's one guy who has a game-theoretic argument, which may or may not be sincere. And it looks like he's persuaded at least 25 other people to go to a rally at some point. But you keep generalising from this one guy to the voting patterns of hundreds of thousands of other people, you talk about "Michigan Muslims" instead of "Hassan Abdel Salam and anyone who happens to be listening to him".
Besides, the whole game theory analysis only makes sense if you assume that Michigan Muslims would "naturally" vote for Democrats. Muslim Americans tend to be socially conservative, working class, religious, and legal immigrants, who would tend to be more natural Trump voters. On the other hand you've got Trump's "Muslim ban" policy of eight years ago which would tend to push them away from him, but that's not a policy this time around. So it's fair to model Muslim voters as people whose preferences are not particularly well aligned with either party.
Muslims in the US are destined to politically walk forever through the valley of the shadow of death, I am afraid. To threaten that “if you do not give me something more, I’ll vote for those who will give me even less” is not a credible threat. No amount of over-thinking can change that.
Maybe they can get something by splitting their vote, though. Something like “unless you give me something more, I’ll only vote for you in presidential elections - in Senate and House of Representative races, I will vote for you opponent. “ Assuming that securing the President is more important for Muslims (Presidents have a lot of leverage in foreign policy), while winning in Senate and House races are less important for them (since that's more domestic policy stuff), but quite important for the Democrats.
It’s not likely to be effective, either, but at least it is something.
The voting dynamics are interesting to think about, but in the real world there are a lot of other payoffs. They're are political appointments, policies, pork, election resources for future races, and probably many others.
If you are helpful in the coalition, then you can jockey for those other prizes. And in the case of the Michigan Muslims, I think they probably had some hope of such a prize from Harris, even if forcing Israel to give up wasn't in the cards, and no hope at all from Trump.
The rage toxo post (aka best SSC post ever) suggested that we hate apostates more than infidels. I think that's what's going on here. They felt that Harris should have been on their side but wasn't, so were way more mad at her than at Trump, who is ideologically far away and therefore there's no sense of betrayal.
My only incremental contribution is to agree with Scott insofar as the fact that this is not an obviously solved question makes it possible that almost anything is a "strategy." This is just a variant of the Folk Theorem of game theory, that almost anything can be a dynamic equilibrium with sufficient patience.
This seems overly complicated. For all the overheated rhetoric, few people actually think that this is going to be the last election. As such, it’s just iterated game theory, and the game being iterated is a sort of ultimatum game. If you think you were given a shitty offer you reject it and eat the short term cost in the hope that in the next round you’ll be offered a better deal.
I'd been thinking about this too (not with Israel, but with policy in general) and ran into two big questions:
1) Shapely Values are notoriously non associative. If I calculate Harris and me, it's reasonable that I get nothing. If I calculate Harris and People Like Me it's more complicated.
2) To what extent should Trump be seen as a threat by the parties and ignored? In a normal election I'd say yes: the shadowy elites that run both parties, while not literally one person, use mental models of each other to act as a single unit and maintain strict solidarity. But Trump is, maybe, not part of that coalition. Though the coalition still decided to back him. So I don't know.
I would think the best approach is to try to coordinate some sort of write-in campaign. The Prophet Muhammad, Yasser Arafat...doesn't really matter who; you're just trying to send a message to both parties that there were votes up for grabs that went to nobody.
Part of it is punishment, sure. I mean that seemed to be the collective strategy: You have taken us for granted and not served our interests, so we are going to punish you.
But an equal or larger part is: You have shown you cannot solve this problem, so we are going to take a (maybe foolish) chance that this other guy can.
> One of the few checks that voters have on the Democrats’ corruption and incompetence levels is to threaten to vote Republican. Is there some level of Democratic corruption at which I should vote Republican to “punish” the incumbent even if I think the Republican would be a worse leader overall? I’m not sure, and currently lean towards no, but I can’t say it doesn’t tempt me.
I really think you're suffering from failing to distinguish trade offers and threats. There are levels of Democratic noncooperation where you vote third party for a candidate who offered you more -- that is, throw away your vote, in the present system, but in an organized way that shows your power to show up at the polls and your responsiveness to offers. It never makes sense to vote for Trump (unless he's made you a better offer). I would expect actual political operatives to treat these bargaining positions differently, just like a small-town market outside of California differently considers the positions "I'm not shopping here if you offer me too few grapes per dollar" and "Gimme more grapes or I'll break a window."
> I really think you're suffering from failing to distinguish trade offers and threats.
I think what you mean by 'threat' is sufficiently different from ordinary senses that it might be useful to invent or repurpose another term to more clearly differentiate what you mean. (One of) the dictionary definition(s) seems to cover what you mean but ordinary usage has expanded far beyond its narrowness. Sadly, I couldn't think of any good and pithy alternatives.
Some not-particularly-satisfying alternatives/qualifications:
Maybe – but it seems perfectly understandable and relevant to other decision theories and game theory too. A non-extortionate threat should be (potentially) simply avoided or its negative consequences mitigated, whereas an extortionate threat should also probably be disincentivized too.
> Trump is left actively better off by his decision to be so hostile to the Muslims
I don't think this is a correct. If Trump were less hostile to Muslims, presumably they would be even more willing to vote for him. So there is no incentive for him to be hostile.
I think a better model here is that one candidate has policies A, C and another B, C. The voters prefer "A and not C", and since "not C" for them is much more important then A, then they flip to the second candidate.
If the voters in question actually liked Trump's other policies, then following the same logic they
should have switched to Kamala to punish Trump for an incorrect stance on the Muslims.
But he's not hostile to Muslims this time around. Eight years ago he was promising to ban visas for all Muslims. This time around he hasn't talked about it. Part of that is a lack of recent Muslim terrorist attacks in the US, but part of it is that everyone knows Muslim votes are up for grabs in Michigan.
We could also look at the other side of the game theory here -- does the threat of Trump's threatened Muslim ban dissuade terrorist attacks in the US?
If you didn't protest-vote in this election, given everything including the Democrats and most of your personal media lying to you for at least months and reasonably likely years about Biden's mental faculties leading to a candidate who nobody wanted in the first place, I don't think you're willing to protest-vote at all.
So, if you acknowledge a protest-vote is valid, I think your first step is noticing that you need to be a lot more willing to protest-vote in the first place.
More generally - if you're a member of a coalition, and whatever interest group you believe yourself to represent isn't doing better as a part of that coalition than if you weren't part of that coalition, that's a good time to protest vote; your interests are being disregarded.
As a more useful metric: If your voting group is being pushed to go out and vote, that is the time to protest vote. If your specific group is being specifically pushed to go out and vote, it's because the party apparatus knows that your group isn't very enthusiastic about the offer being made and probably won't show up to vote in their usual numbers, which probably means that your group is being offered a raw deal by the coalition.
So, for example, the Democrats were really pushing men to go out and vote. Well, the party is still operating as if it is the early 2010s and it's socially acceptable to shit on men while demanding they support women (Christ, I haven't seen so much use of the phrase "toxic masculinity" in nearly a decade). Unsurprisingly, men turned up and voted against them, because by and large we're done with this bullshit. I'd hazard a guess that the misandry inherent in the election was one of the key mistakes of the effective campaign.
Interesting post. I had a similar thought a few days ago ( https://www.griffinknight.com/p/voting-to-send-a-message ). I don't really go into the game theory of it, but I tried to think through some sort of framework for when "voting to send a message" is a viable strategy in that it makes things "better" for you (of course what is better is subjective based on the voter).
I think the right questions to ask oneself when protest voting are: 1) is either outcome of the vote relatively unimportant and non-consequential to me? And 2) am I confident that my intended message will actually be received and actionable? If the answer to both of these is not an astounding YES, it probably just makes sense to vote for whichever direct outcome you think will be "better" for you.
The limitation of this kind of analysis is that it treats politics as a constrained optimisation problem in a simple closed world. But strategy does not exist in a closed world. Finding a good strategy is a problem requiring creativity. Even in something as simple as the micro mouse racing competition, creativity dominates optimisation.
The real world is overwhelmingly complex. A good strategy comes first from a diagnosis that reduces this complexity, and a guiding policy that governs how you act. If you know everyone's reduction, perhaps this is simple enough to do constained optimization.
It's true that Democrats reduction was leading them to ignore the Muslim vote, and that trump is worse. But that meant Muslims had one choice: try to get democrats to think again, to move outside the current paradigm. They failed, but that doesn't mean it was wrong to attempt.
The Democratic policy was based around the idea that they should avoid responsibility for Gaza. Give Israel all the weapons to satisfy the Israeli lobby, make gestures towards humanitarian concerns for the Muslim vote. Don't attempt to look for other solutions, lest we get the blame.
But this ignores that the Israeli governments actions are bad in the long term for Israel, the US, and the Jewish diaspora; and the Israeli public support it largely because they have lost hope. Any strategy that stops the ethnic cleansing in Gaza must involve convincing these constituencies that a better way is possible, but that won't happen if all you are prepared to offer is weapons and hand-wringing.
I think that most of these analyses don’t adequately consider the long-term consequences of the relationship that is being formed …
In a martial contest, if person A does not in some way threaten person B, then B will eventually wear down A with a series of attacks and A will always lose. Even when defending, a person must show the threat of a counter-attack or they are doomed. There is a 2500 year history behind this theory, e.g., Sun Tzu.
So unless the Michigan Muslims want to trust in Harris’s commitment to “do the right thing”, they must offer a credible threat. If Harris does not in some way fear the Michigan Muslims then she will continually give away pieces of “their pie” and they will be left with a total loss. Since “do nothing” creates a long term certainty of total loss, and absent any other potential courses of action, their action seems well chosen.
Once the opponents are balanced then negotiations can start.
Not going to comment on the Muslim issue specifically, but with some factions on the left, there’s a real problem of them taking any compromise and using that as the basis to demand more concessions. After a few rounds of that, those factions lose all influence.
I've heard that Brexit happened that way. That a solid majority was against Brexit, but many of them decided to vote for it anyway to signal disagreement with David Camerons policies.
I don't know whether it's accurate. There were certainly protest voters like that, but I have not investigated whether it was really unusually many .
I don't think that was really true. What did happen is that there was nothing that forced the Brexit advocates to define their offer in concrete terms, so they could portray it as the answer to whatever you were irritated about -all things to all people. So they appealed to both people who wanted protection from the international markets and people who wanted less. Social conservative protectionists and free-traders.
Remain had a narrow 52-48 lead in the polls, which turned into a 48-52 loss when votes were actually counted. This is easy to explain with ordinary systematic polling errors (shy Tories and the like).
In a real election, I'm skeptical of any of the probabilistic stratagems. Even if they are perfectly executed, the candidates can never be sure that they actually _were_ executed.
Hmm... Re the Electoral College, perhaps it is time for a coalition of disempowered non-swing-state voters to organize for a popular vote...
An alternative is a coordinated vote for someone else, a third-party candidate or fictional character, especially when the person to vote for is arbitrarily decided in public shortly before the election. Just like protest marches, they convey the unspoken message "I can mobilize a whole lot of people who do what I want them to do, therefore I have intent and means, and therefore I am someone worth negotiating with".
For fun, take a look at this webcomic; page 87 is where I get the "intent and means" language from.
>An alternative is a coordinated vote for someone else, a third-party candidate or fictional character, especially when the person to vote for is arbitrarily decided in public shortly before the election.
Yes, that would convey power, but also note that it is a deterministic action. For any of the _probabilistic_ strategies to be credible, not only would the leader(s) have to coordinate the vote, they would have to publicly declare the strategy, publicly show the resolution of the random part, and this is on top of the actions you describe. I think it would be too hard to do. The deterministic part, by itself, is already hard to do.
Yeah, probabilistic is not so good when performing for a human audience. People like feeling as though their leaders have a plan, and we don't all have IQs of 130.
Talking about taboos and savvy brings us out of game theory, but since you brought it up...if your block never breaks taboos but just relies on being savvy and accepting handouts proportional to power, you risk being rolled by people who actually are savvy and trick you into believing you don't deserve much.
Is this what happened to the black vote? It used to be reliably Republican, then some of them started saying "actually the Republicans aren't satisfying our interests enough, we're going to vote for the Democrats despite the slavery and Jim Crow and shit, wink wink". But enough of them didn't hear the "wink wink" that they went and became a reliable Democrat voting bloc, whose interests still aren't being satisfied.
I think it was LBJ pushing through the Civil Rights legislation that moved blacks toward the Ds. And then the local white backlash was embraced by Rs, which further moved blacks toward the Ds.
One problem is that vote totals are public. So politicians will look back and see "not many folks here voted for Trump", and assume that you're irrelevant.
Eliezer Yudkowsky is not actually an authority on game theory. His interpretation that claims Logical Decision Theory isn't just bunk is not supported by real experts. I like a lot of his work but please don't treat him as a good game theory consultant instead of an actual economics professor.
To he clear, Eliezer is a published reseaecher in this branch of philosophy/mathematics.
More to the point, if his arguments are wrong, it should be possible to articulate *why*. (And then once you do that, the argument from authority becomes irrelevant to anyone with the time, patience, and intelligence to judge the debate!)
1) Like he says in his arguments, he has not actually mathematically proven his research, "First of all, because nobody has actually derived the LDT Ultimatum solution from first principles.", he's just going off vibes that he thinks his conclusions should be true
2) If LDT works, then you should be able to set up simulations/experiments where people/bots behave according to LDT principles. You can create all sorts of experiments with lots of bots operating under different strategies play Prisoner's Dillemas or other games, and the winners are consistently what CDT predicts they'd be. When humans play, sometimes results differe, e.g humans rejecting "unfair" offers in the ultimatum game, but I think that's simply explained by humans feeling like reputation matters even when in the rules of that specific game it doesn't.
3) I think the fact that mainstream economics doesn't agree with him, on top of the fact he admits he hasn't rigorously proven LDT, plus the fact there's no experimentalevidence supporting it, adds up to a very strong case that it's bunk.
It's certainly an unusual strategy, which means that either they're missing something or every other interest group is (or there's something else unusual going on in this specific case, but I have no idea what that would be). "Not voting at all" seems like something that other interest groups might do and feels a lot less extreme, even though it's not really that dissimilar from half as many people voting for the opposition instead.
Maybe they're too new of a group to have other forms of influence? In my model of, say, a teacher's union, if they're unhappy with the local Dem politician, they might fund a primary opponent, organize protests or campaigns to call/write their office, write letters to the local newspaper, threaten a strike, etc. I don't know if they thought these were not viable options. Overall though, it seems like the strongest incentive is the one that says "we don't negotiate with... defectors" and just ignores them.
I think there might be too much mathing going on here and not enough thought about human nature around personal wants and tradeoffs.
Has anyone ever run a version of the Ultimatum Game with an actually important amount of *real* money on offer? Because I think the official studies made a big mistake in offering trivial amounts of cash.
Ten or fifty or even a hundred dollars has almost no value to most people, and thus if the percentage offered is insultingly small, person B might very happily consider "losing" $0.01 or $2.50 or $10 a good value to punish the other person for their "greed."
But if the amount of actual real money on offer is $100,000, "unfair" split offers of 90/10 or even 99/1 are going to be a lot more tempting for person B!
I know I'm not willing to forgo $1000 merely to prevent someone else from having $99,000. And I'd be *delighted* to have $10,000 free dollars for doing nothing except saying, "sure, keep $90k."
If the amount of real money on offer is $1,000,000, I'm willing to take an even smaller "insulting" percentage.
I have a strong intuition that the distinction *REALLY* matters. The Ultimatum Game doesn't show us anything except that people will cheerfully spend trivial amounts of money to punish others because they see it as a good value.
This may indicate that the Muslims didn't think Harris or the Democrats ever had anything "real" to offer them, and that they might as well punish them for not offering / delivering more.
I asked pretty much the same thing you did in a graduate seminar I took almost 20 years ago, with the same thinking behind it. IIRC, the professor's answer was that there are studies done with medium sized prizes (on the order of hundreds of dollars), with test subjects living in poor countries for whom the prize was quite significant relative to their incomes. In those studies, people were still willing to reject very unequal divisions, but were more apt to accept moderately unequal divisions than the same pool of subjects were with smaller prize pools.
Huh. Would those studies have been able to control for more reputation/honor-based cultures, which many poor countries are likely to have? So that there was a greater perceived value in saying "fuck you" to an insultingly low offer?
But more importantly, was the prize amount in the poor countries actually *REALLY* significant, like, buy-a-car or launch-your-kid-to-a-better-life or never-go-hungry-again significant?
$1000 is a lot of money to me, but I still *might* say "fuck you" to someone offering it out of a million.
But if I was offered $100,000 out of $10,000,000 to pay off my mortgage, I would certainly take that deal. I would take $100,000 no matter *what* the deal was. There's no amount of "fuck you" satisfaction that would be more satisfying than having $100,000.
It's hard to imagine other people not agreeing when the money is really, *really,* real, but it's certainly not the first time I've been wildly wrong modeling other people!
That, however, would rule out human nature as a linear system of equations, which would make it much more inconvenient for academic researchers to "solve". So we can't have that.
Also, good luck getting a large enough research grant to allow doing million-dollar ultimatum games enough times to draw statistically useful conclusions.
I think the controls were running the game with different prize pools amounts with participants from the same culture, which seems like it should have accounted for different cultural attitudes towards stuff like honor, collectivism, egalitarianism, and hierarchy which might have affected willingness to accept unequal divisions.
I think the dollar amount was in the range of a few days or a couple weeks wages, not "enough to retire on" and definitely not "generational wealth". I agree with your expectations that for the much larger amounts, it should converge on accepting so long as the absolute amount offered is enough to really turn your head regardless of the total amount in the pool.
I would guess that the larger the pot of money on offer, the less likely that the game will be repeated. The sensible strategy (both game theoretic and instinctual) is very different for iterated game theory vs one-off games.
From the abstract: ”rejections were less frequent the higher the stakes, and proposals in the high stakes conditions declined slowly as subjects gained experience.”
It’s interesting. If you look at figure 1c, you start seeing an increase in 33% offers as the stakes grow and players learn - and on the responder side, everyone stops rejecting these offers.
Why Muslims? Why not small business owners, or conservative Christians?
This sort of strategy only works if you're a group that has a reputation as angrily and self-destructively defecting if you don't get your way. The kind of people who will start wars they can't possibly win with their more powerful and better armed neighbours. The kind of people who will blow themselves up in crowded marketplaces for the joy of killing a handful of enemy children. The kind of people who'll fly passenger jets into skyscrapers just to gain a reputation as a group that shouldn't be fucked with, knowing that it will lead to a decades-long war that kills millions.
This kind of self-destructive defection is very much on-brand for Muslims and very much off-brand for everyone else. I would say that it often works out well for them, but that overall the costs they repeatedly incur from the many defections probably outweigh the benefits, which is why half the Middle East is rubble.
It was confounding to see that all sides on the Israel/Hamas/Gaza war punished Harris and the Democrats and seemed to drift towards Trump. First blush would be to think the (D) positions and messaging were uniquely bad and contradictory and self-sabotaging. But the Harris campaign said nothing about anything, and Trump also said very little recently re: the Middle East.
I conclude that everybody is unhappy (to put it almost grotesquely mildly) with the war, and the party at the helm would be identified, even subconsciously, with the misery. It was probably the Right Move by all parties to say nothing much. Because what could one promise that people would believe? Whom could you champion that would not "Other" the rest of the electorate?
As Scott has noted, Trump has been a supporter of Israel and the Israeli government through all of his public life. Maybe there comes a time in conflicts where *certainty* is more valuable than specific outcomes. Maybe any sort of concrete action is better for all sides than chaos. Certainly Joe Biden, still US President, was not going to make any decisions. It is a strange result that Donald Trump of all people presents more gravitas to the world that the current US policymakers.
A quote for which I cannot remember the attribution, but can never forget the implication:
"...can't keep his feet out of the sh*t on both sides because he can't get the fencepost out of his a$$!"
Well I think there’s a strong case to be made that if Trump was president, Hamas would not have done 10/7 and so there wouldn’t be thousands of dead Gazans. Hamas strategy is based on the US not giving Israel a blank check to crush them since they can’t beat Israel militarily.
I think this line of thought about what Michigan Muslims should do is completely wrong-headed. Most of the US is now so angry that people have no intellect left over to think about the pros and cons of voting strategy. The are desperate to scratch the anger itch, to get rid of the feeling that fuckwads who are dumb and wrong are doing things they hate and they have no power to stop them. Of course the present election is especially toxic and contentious -- though that may not change going forward, because a lot of the monstrous rage is created then aged in ideal conditions on social media, like monstrous wheel of limburger, . But in all situations where there are conflicting interests it's highly likely that some will be so overtaken by anger delusions that they are incapable of thinking about tactics, likelihood of various things, long term consequences of this and that, etc.
And the present state of ACX commenters is excellent evidence that rationalism does not protect people from having their mind overtaken by primitive anger, in which obviously ridiculous ideas seem true: Everyone who voted against my candidate is either evil slime or a moron. it is fine to mock them, insult them, etc. All of my opinions about my candidate are well-founded. I am able to make an accurate forecast of how things will go if candidate A is elected, and if candidate B is.
Not everyone who posts sounds like they have a case of that, but many do. Scott just gave somebody posting on the prediction market threat a brief ban, commenting that it's brief because everyone is crazy right now.
There's a book I'd like to recommend. It's the best corrective I've found for my own tendency to slide into Rage Idiot mode. It's *How to Think,* by philosophy prof Alan Jacobs. A better title would be *How to be Fair-minded*. This guy really understands the routes by which people end up believing and acting on toxic nonsense. He sees the process as driven mostly by efforts to manage self-esteem in situations involving ambition, desire to affiliate with an appealing group, fear of ending up in the lame group. .So he focuses on the emotions that are stirred in these situations, not on what constitutes clear thinking. I highly recommend it as a way of improving one's judgment and self-management in conflict situations.
First, framing this as Michigan's Muslims "defecting" from the Democratic coalition, is predjudicial. Part of what the Democratic coalition has traditionally offered, and part of what made Michigan's Muslims join that coalition, was a commitment to human rights for oppressed peoples around the world. At very least, to e.g. complain to the United Nations when someone tries to perpetrate a genocide anywhere. And probably most Democrats have that something like thirty-five levels down from the top of their list of priorities, but presumably the Detroit Muslims care about the human rights of the Gaza Palestinians rather more than that.
From their point of view, it's the Biden Administration that has defected from the Democratic coalition, by basically signing off on fifty thousand dead Gaza Palestinians. And maybe that's the right thing for the administration and the party to do, for moral or geopolitical or domestic-political reasons. Sometimes coalitions have to cut loose some of their members to better serve the interests of the rest. But that doesn't mean it's wrong for the Detroit Muslims to believe that they are the wronged party here.
And as for what to do when you perceive that the other party has defected, I think that's pretty much straight out of Game Theory 101. If it's a single contest, you take the L and salvage whatever you can out of the matter, take the $0.01 in the Ultimatum Game or whatever. If it's an iterated game, you punish the defection and then renegotiate for the next round.
If Donald Trump is going to egg Benjamin Netanyahu on to literally exterminating the Palestinians, then this round is all there is, and you vote for the politician who just *might* try to stop that and will at least probably open the doors to more refugees than would Trump. If the present war is going to kill say 4% of the Gaza Palestinians and then we'll have to deal with the reconstruction after this war, and then the next war and the reconstruction after that, then maybe it's worth enduring four years of Trump for a chance at having a reforged Democratic coalition that takes their concerns seriously going forward.
And it's tempting to extrapolate from there to "see, this proves that all those whiny activists don't *really* believe all that genocide nonsense!". But really, all those whiny activists are neurotypical human beings rather than nerdy rationalists, so they don't do the math for Game Theory 101. They just implement long-established heuristics, which are basically "everything is iterative, punish all defectors, we've hardwired your brain so it feels *good* when you punish defectors, and surely you can come up with the case-specific rationalization".
> the longest Discord thread in the history of Project Lawful
Good times. :-)
And yes, it's important to understand your goals. Are you trying to achieve an equitable balance of power in a coalition, regardless of how little actual power that might give you? Or are you trying to maximize your short term power in the coalition at the expense of long-term interests like avoiding backlash? (One must note that, in the case of Palestinians, there's a war going on.)
I think a good choice is to demonstrate power and control by cementing a voting block that acts in a coordinated way. Announce the day before that your people will vote for "Caliph Haroun El Poussah", or Cornel West, or someone else neutral who won't win, and then let the vote totals show the outcome. Eventually someone will want to woo your block, or not. Another complication is that that's negotiating in good faith, which is a quokka-like trait that may not have any place in politics, even if an ASI might appreciate it.
As for how much to ask for, that's where the Shapley value comes in, but like you say, it's important to remember that the coalition involves a lot of other agents. If your goal conflicts with the goals of more important coalition members, then you're kinda screwed. On the other hand, it's sometimes possible to negotiate directly with other coalition members, and get them to change some of their goals that are less important to them, to match yours, while you adopt some of the goals important to them. (As suggested in Eliezer's "Three Worlds Collide".)
Harking back to the "threat" discussion, it's dangerous to take actions that harm your own interests, as a minor player in a multi-agent system. If there's an R coalition and a D coalition, and you want to join the Ds, and the Rs aren't ever going to help you, there's always a chance that harming your interests and the D's interests will result in an R victory. The D coalition, being bigger and stronger, is probably better placed to recover than you are. (Because it's not just a competition for abstract points; political losses can actually damage your ability to act in the future.) Plus it pisses people off, and they may decide to let you keep hurting yourselves.
Personally, I usually take a low-brainpower way out. In races where I don't actively support one candidate, and where it's predicted to be a blowout, I vote for the underdog. That makes the race more competitive. Now that I've articulated this in public, though, I should point out that Kant's Categorical Imperative has some things to say about coordination, and so I wouldn't recommend that the entire world blindly adopt this low-brainpower policy.
- and that motivation is that they want to punish it.
Both of these strike me as not obviously true. They are what the Dems want to be true because that fits into the rest if their theory of the world, and doesn’t require them to rethink anything, but that doesn’t actually make them true…
I’d put the Arabs more in the same bucket as Hispanics — utterly sick of being defined by their ethnicity above all else, and to add insult to injury, having that ethnicity defined by people who thinking reading Said makes them experts on how you think.
"Is there some level of Democratic corruption at which I should vote Republican to “punish” the incumbent even if I think the Republican would be a worse leader overall? I’m not sure, and currently lean towards no, but I can’t say it doesn’t tempt me."
There should be, assuming the Republican has good character (i.e. a John McCain type Republican, who has good character but horrible (to me) policy). Because a person of good character will research, experiment, and reassess to ensure best outcomes are met, not be blindly ideological. For example, this is why Pete Buttigieg was such a great Secretary of Transportation: he kept looking for the best ideas to achieve his goals, and found them, even when they were against his initial positions.
When I see someone say: "Let's all vote in X way, and I acknowledge that this looks like it's contrary to our interests but here's my argument for why we should do it anyway," my first thought is to check whether the speaker is actually a member of my coalition (as opposed to a member of the other side engaged in creative vote-gathering).
The biographies at abandonharris.com look pretty real to me, so in this case I'm not seriously suggesting this was a trick. But I do have to wonder how long we have before it becomes possible to generate a "Movement of X Interest Group Voting For Y Candidate" page algorithmically.
I think voting third party for someone you really would like to be president more is the answer here; it maximally hurts Kamala, but since you're not voting to hurt yourself, I don't think it counts as a threat.
Anyway, perhaps these Muslim voters don't want dead Gazans more than anything else. I think Hamas would have been too scared to do October 7 with a Republican president. So, no dead Gazans.
I think trying to apply analysis from perfect-information games with rational players to politics is generally a mistake, and that working out a correct strategy probably involves a deeper understanding of the psychology not just of the people you're ultimately trying to influence, but of the people in your own movement that you're trying to persuade to adopt your preferred tactic, than most people - definitely including me! - have.
The game theoretic approach only works if the other players also follow game theory. eg: beginners luck in poker.
Politics works on shock value. 'The muslims voted for Trump' drives better shock value than any of the alternatives. Therefore, you vote for Trump. Plus, 1 vote is a 2 vote swing if it goes Kamala -> Trump, but a 1 vote swing from Kamala -> 3rd candidate. Always maximize your power. Go Kamala -> Trump.
However, shock value only serves to draw attention. And attention isn't always good. Attention helps highlight a show of power, only IF you have power.
The way the math played out this election, PA was always going to decide the outcome. If Trump won PA, none of the other swing states added up for Dems. If Kamala won PA, she could lose 1 swing state and would've likely swept the rest. Plus, Trump & Obama have showed that people vote based on a national vibe. Since the 90s, PA-Wisc-Mich have all swung in the same direction. The mid-west [1] / great-lakes mega-region [2] swings together.
From the attention drawn by Dearborn, Dems are going to realize that American Muslims don't matter. This time, Dems sat on the fence by allowing campus protests to get really bad and rejecting Shapiro as the obvious swing-state VP candidate. Clearly, it wasn't enough for Muslims.
Non-black (NB = not descendants of slaves) Muslims are 1 % of the US population, and Michigan is the only swing state with a substantial NB Muslim population (2%). So at best, a perfect pro-democrat Muslim vote base can help make a 2% difference in Michigan. That's pretty small. Especially when it is traded off against the most important state (PA), which has 3.5% Jews in comparison.
Going off attention and power as the 2 metrics of interest. Here's a better decision tree for protest voting an already lost election :
1. If in_state_power can flip future national election -> protest vote such that it draws most attention = vote Trump.
2. If in_state_power cannot flip national election -> vote to create narrative around power and minimize attention = vote Dems.
In scenario 2, all michigan muslims vote Democrat. At a 1.4% difference, it is a swing-able number for NB michigan muslims (2%). Kamala still loses nationally, but she wins 1 and only 1 swing state -> Michigan. New narrative : The country abandons Kamala, but the Muslims pull out unexpected win.
Now the same 1.4% creates a greater impression of power in this new narrative. The narrative doesn't get much scrutiny because it falls in line with existing preconceptions (muslims matter, muslims vote dem). In 2024, the Dems gave NB-muslims more leeway than their demographics demand. NB-Muslims might think that this was not enough for them, but their inflated entitlement doesn't match their demographic power. The best thing to do, is to let the status quo continue and lay low. I predict, they will regret drawing attention to themselves.
I'm getting ahead of myself, but 2028 will be about the collective will of the 2 main mega-demographics: the great-lakes mega-region & Latinos. If you aren't part of either mega-demographic, you don't matter.
I did a lot of ruminating on this back in the day, when I was a little bit more active trying to help get gay rights legislation passed in the US. The DNC was *awful* about actually putting their own resources behind advancing gay rights, to the point that Clinton in the 90s was actually signing anti-gay laws like Don't Ask, Don't Tell. But obviously we still had to keep voting for them, because even though they were merely the ones punching us slightly less, the alternative was "actually homosexuality should be illegal". Even when it started being a wedge issue on the other side, such that running on gay rights got them more votes than it cost, they still basically let the courts do most of the heavy lifting. To this day, very few laws that you could call "pro-gay" have ever gone through the federal congress, most of them basically symbolic.
But unfortunately, I don't think there's really a viable alternative to "vote for the guys passing slightly fewer laws banning your community from public life, then try to route around formal politics and hope you can change the culture." I mean, it's left me feeling permanently rather burned by politics, but I don't think the DNC has to actually care about that? Without some viable BATNA, they can suck as much as they want. As long as there's an effective duopoly on governance, these parties pretty much have us by the short hairs.
if we're not counting the effort of going to vote, I don't get yud recommending abstaining or cancelling out votes as opposed to voting for trump. it's votes that you add together, so voting trump seems like doing the same thing but harder. the isThreat(x) function is sus here
I asked Eliezer this same question when I talked to him. He said:
"I agree that these two impacts are similar. Similarly, if you're negotiating for an artwork with somebody who values their life at a billion dollars, the impact of "I'm gonna refuse to make this trade which has a gain-from-trade to you of $1M with 10% probability" and "I'mma shoot you with 0.01% probability" are the same to them. But the decision theory surrounding it is not the same."
"But the decision theory surrounding it is not the same."
That sounds completely intuitive to me, but does anyone here know if there's somewhere Yudkowsky has elucidated this at greater length?
Unironically, it's explained very well in planecrash if you don't mind reading millions of words of BDSM-themed Pathfinder roleplay with in-world maths lectures.
I don't think we put out any explicit paper on the Nature of Extortion because of the number of unsolved problems in it, combined with a broken journal system that wants a pretense of full solutions.
Basic intro to LDT at https://arbital.com/p/logical_dt/?l=5kv.
For some simple ideas going past pure intuition of why "the decision theory is not the same", consider such differentiating observations as:
- In positive-sum bargaining, the other party would rather the whole interaction occur because they have a chance to profit; in extortion, they'd rather the whole interaction not occur. So any sort of coordination work, overhead, or decision-theoretic posture that helps the interaction go through, the other party is motivated to do for negotiations but not for threats.
- A rational agent should always do at least as well for itself as a rock, unless it's up against some other agent that specifically wants to punish particular decision algorithms and will pay costs itself to do that; just doing what a rock does isn't very expensive or complicated, so a rational agent which isn't doing better than a rock should just behave like a rock instead. An agent benefits from building into itself a capacity to respond to positive-sum trade offers; it doesn't benefit from building into itself a capacity to respond to threats.
- Consider the Nuclear Prisoner's Dilemma, in which as well as Cooperate and Defect there's a third option called Nuke, which if either player presses it causes both players to get (-100, -100). Suppose that both players are programs each allowed to look at each other's source code (a la our paper "Robust Cooperation in the Prisoner's Dilemma"), or political players with track records of doing what they say. If you're up against a naive counterparty, you can threaten to press Nuke unless the opponent presses Cooperate (in which case you press Defect). But you'd have no reason to ever press Nuke if you were facing a rock; the only reason you'd ever set up a strategy of conditionally pressing Nuke is because of a prediction about how your opponent would respond in a complicated way to that strategy by their pressing Cooperate (even though you would then press Defect, and they'd know that). So a rational agent does not want to build into itself the capacity to respond to threats of Nuke by choosing Cooperate (against Defect); it would rather be a rock. It does want to build into itself a capacity to move from Defect-Defect to Cooperate-Cooperate, if both programs know the other's code, or two entities with track records can negotiate.
- If Harris had no more ability to respond to threats than a rock, you might offer her your vote conditional on her making a sufficient offer to you, but you wouldn't bother to threaten her with setting off a nuke / voting for Trump (hurting her at cost to yourself), because she wouldn't have built the capacity to respond into herself. An ideal rational agent doesn't want to build a capability to respond to threats to vote for Trump, because it doesn't want to be part of that kind of interaction in the first place; it would rather be a rock in this regard. It does benefit from building into itself a capability to respond to positive-sum trade offers, like offers to vote for it in exchange for its pro-Muslim policy.
As always, do not misunderstand LDT/FDT as making any sort of "commitments" or "precommitments"; the whole point and idea of LDT/FDT is to compute the correct course of action in one sweep via evaluating expected utility using more sensible counterfactuals. It is only other decision theories and viewpoints that need to imagine talk of "commitments" (the places where that other decision theory would rather behave differently than it usually would) in order to see what LDT computes naturally and in one shot. LDT doesn't need to "commit" to paying off the driver in Parfit's Hitchhiker; it just uses more sensible counterfactuals to compute the superior result of paying off the driver.
Thank you! It sounds like you're saying the asymmetry lies in that threats would only be made in the first place if the perpetrator expects the victim's decision procedure to have some pathological flaw that would dispose them to give into such threats, whereas trades would be offered even if the counterparty has a perfect decision procedure.
Your ideal decision theory might not need precommitments, but humans certainly do and can benefit from them.
So there should be something in a practical decision theory that compiles to precommitments, when implemented and executed by mere mortals?
I am not sure that Cooperate : Defect : Nuke is entirely isomorphic to Vote Harris : Abstain : Vote Trump.
In my opinion, the Muslim's dilemma assumes that absent any coordination, it would be in the interests of the Muslim block to spend the costs to vote for Harris because the benefits of a pro-Israel Harris presidency over a Trump presidency.
If that is the case, then them voting for Harris out of self-interest is the baseline for non-coordinating actors following their self-interest, and any other choice by them would be akin to the nuke option (hurting themselves to also hurt the other player in an effort to gain better outcomes in counterfactual situations), aka as a threat.
The only difference is that there is some cost to the act of voting itself, which is presumably small.
In general, I am still confused about what constitutes a threat that rational actors should ignore and what constitutes a reasonable response to encourage better outcomes in counterfactual situation.
For example, not taking unfair splits of 10$ with some probability seems to be equivalent to the nuke option: you burn some of your utility in order to frustrate the utility function of the other player.
From the payoff matrix alone, you can not determine that a certain split, like 5$ : 5$, is objectively fair. Perhaps both parties have to put an equal amount of work into it and are equally likely to find themselves on either side of the table. But I can also make up a story with equivalent payoffs where one party is an honest trader with a cartload of silks and the other one is a robber (or medieval customs official) who threatens to burn the silks unless he gets a cut.
And in the real world, things are even messier. Should we assume that the utility of 5$ is the same for everyone, or assume that it just increases the net worth of the recipient, and is subject to some concave (possibly logarithmic) utility function? How does the good-will of the pro-Israel lobby (which Harris presumably wants) compare to the desire of Muslims to minimize the deaths of Palestinians? Neither one can easily be translated to dollars.
Maybe tangiental, but those aren't quite equivalent for two reasons. First is almost patronizing to point out, but money doesn't have linear utility, so its exchange ratio with other goods can vary with exchange size. It's under-defined what "valuing your life at $X" actually means, but it'd be reasonable to take the deal "Y% chance of dying for Y% of $X" for some values of Y and not others.
But also here's a broader point about "value of life." I think in this community it's often viewed as "amount you'd be willing to exchange your life for," which implies you put value on money even after death (maybe because you wish to donate it, or have family, etc). But as I understand, it's computed by looking at how much (usually small) risk people are willing to take in exchange for some amount of money. In that case, the point of the money is to use it in the many worlds which you're left living. So those numbers really only make sense in the low-risk regime. Even if utility was linear with money, it would be consistent to be play Russian roulette for $10M, but not kill yourself for $1B, because 5/6 of the time you get to enjoy your money in the first example versus 0 in the second.
Maybe a correction for point #2 is that people have two utility curves for money, one if they expect to be alive to spend it and one if they don't. That I think knowing both would be enough to figure out what risk deals a person would be willing to take.
> In that case, the point of the money is to use it in the many worlds which you're left living.
Not necessarily. Faust sold his soul to the devil in exchange for the benefits he got for the years before he had to deliver.
It's doing something with the same basic effect but different decision-theoretic implications. It's like how if the Vikings demand £1m Danegeld, and your other option is spend £1m to repel them by force, the basic costs to you are identical but the decision theory is very different (you shouldn't give people an incentive to threaten you.)
I'm similarly frustrated by how many states are captured by one party. I also want to vote on a mix of policy preferences, governing competence, leadership, charisma, and integrity. Berkson's paradox grants us at most one of these.
I can imagine systems that might help but would require unpopular constitutional overhauls.
I have decided to vote against the dominant party in a solid state with a clean conscience. I expect that the dominant party should recognize what's happening well before I actually tip an election. You're more likely to cast the vote that makes them think "ooh, that is getting close" than to tip the election. If there are 70/30 splits, you're not voting for the other party, you're voting for more meaningful elections.
This changes by some escalating amount as things get closer. We all kind of know this intuitively, it's why vote trading appears positive sum, votes matter more (read: at all) in close races. If your race is not close, invest in it becoming so. Or, yes, apply a declining probability of doing so as it approaches 50/50.
Actually if you're serious about this don't just vote contra normative, put up a billboard urging others to do so as well. A single vote resembles buying a lottery ticket and pledging the winnings to your chosen party. ("Don't defect," they say in lottery elections world, "it only works if everybody does it...")
Unless you believe the majority "bad governance good policies" party in your particular state will take the wrong lessons, adopting bad policies while retaining bad governance... that is a risk!
There's also an empirical question about whether parties even act on defecting constituents at all, or whether they are goldfish brained random walks. Immediately after 2000 did third parties become more politically powerful or more ostracized? I think further ostracized... rough news for their nash equilibrium.
If you actually want to impact party decision makers, you probably need to move up the chain and do it through media or money. The mechanisms through which Biden was replaced provide an instructive case study. It involved a somewhat quiet and conditional donor revolt backed by arguments that this was actually the right call, with levers of gradual escalation by making it more public over time.
Can you spark a donor revolt over any of the things you find distasteful? How would you position yourself to do so? (Once you are positioned to do so you will probably find the weapon cannot fire repeatedly and you are stuck in a future resembling the one you were already headed toward without all that effort..)
Arguably though the initial seed on replacement was planted by Nate Silver, so punditry is possibly the most sustainable a vehicle here, even though a longshot much of the time.
Sorry for the threaded reply, thought I was up a level.
Dunno about Eliezer, but to answer why voting 50/50 and voting Trump are different: US presidents are still elected by a minority of the country. Elections aren't won by convincing the other side you're worth voting for; they're won by convincing your own group to actually get off their asses and cast a ballot. If 1,000 people vote 50/50, that doesn't sway the election in the direction of the opponent, but it demonstrates that by acceding, you can get 500 fewer people voting for the opponent *and* 500 more people voting for you, who would otherwise be staying home because they don't want to vote for someone who goes against their core values. And since it doesn't *actually* sway the election in favor of the opponent (i.e., it doesn't on net directly go against their interest), it's much easier to convince your otherwise-preferred-candidate that it's a threat your group will be willing to follow through on than having to convince those same 1,000 or even 500 people to vote for the opponent without a counter-balancing trade.
I feel like framing them as 'defecting' is extremely unfair. They were offered an unacceptable deal from their point of view. Despite being a reliable voting block in the past their concerns were not given much weight.
People being willing to 'swap sides' provides good incentives to those in power. Even if the new guys are worse, if you totally screw us over we might side with them...
It is as though you didn't read the post
If we frame the election as a process that collects information about leader preferences and then selects the best leader on the basis of said information, then we might think there is a obligation under the social contract to feed accurate (to the best of one's knowledge) information to the system. If so, then the aforementioned Muslim protest voters are defecting not from Harris, but from democracy; they are behaving unethically because they're engaging in deception.
(compare being paid to vote for someone, or being paid to write a glowing letter of recommendation for a terrible employee applying to a socially important position, or writing said letter just for the lulz)
SHOULD we frame it that way? I don't know, but my impression is that political scientists tend to see protest voting as caused by inadequacies in the information collection system (e.g. maybe voters are forced to vote for the lesser evil under FPTP but they would be able to vote their conscience under RCV). So here protest voters aren't deceiving anyone; they're communicating imperfectly accurate information to a system that doesn't let them articulate it better. I don't know which inadequacies of this kind would cause the Muslim Trump protest vote; if there aren't any then this is a substantially different case.
Not sure if it makes sense to bring up ethics under a game theory post but maybe this is a good way to understand why someone would disapprove of their choice despite understanding their concerns.
The choice of words might be a bit unfortunate, but 'defect' is just the usual term used in discussions of the prisoner's dilemma.
Honest question: what % of Muslims in Michigan would explicitly condemn the recent attacks on Jews in Amsterdam? https://en.wikipedia.org/wiki/November_2024_Amsterdam_attacks
I’ve never met a Muslim from Michigan so I have no idea. Would anyone here have the answer?
This feels pretty gross and off-topic, and also, who could possibly answer it? I lived in Michigan most of my life and not one of the Muslims I knew was in favor of terrorism. Is that useful information to update your 'curiosity'?
Why is it gross? It could be answered with 'Based off knowing people in Michigan, 99% would condemn this attack. Lizardman factor always exists but 99% of them are very decent people' [hypothetically!]
Why is it not a fair question?
I think it's not a fair question because nobody could possibly claim to know a representative sample of Muslims in Michigan with any real authority, and I think it's gross because it implies it's totally reasonable to assume there might be a ton of people living in Dearborn who are in favor of terrorism.
And, as I said, it's not really on-topic for this post - the post is about whether it's reasonable for people who are concerned about the fate of Gaza residents to effectively support Trump to increase their political leverage, and your response to that was to ask how many of them are pro-terrorism.
All discussions here have spin-off discussions that are pretty far from the original Scott post. And I think it's OK for people to ask gross questions here, questions that imply that a certain group of people is dumb, genetically violent, whatever. Or questions that come right out and say this or that group is savage and inferior. There are enough people committed to "where's the evidence?" that there's no need to feel concerned that some somebody's implied negative view is going to plan the seeds of irrational prejudice in a bunch of readers.
Eh, maybe. My feeling was it was a low-quality comment in terms of how inflammatory it was vs. how effectively-impossible it would be for anybody to give a genuinely meaningful answer to it. I'm not asserting it violates any rules, I'm saying I expect it to drag down the discourse into a bunch of competing anecdotes that make people angry without enlightening anything.
I would like to see different groups get maligned as violent and unintelligent from time to time, it’s gets boring.
Been lots of people on open threads making the case that blacks are dumb and savage. And people routinely say devaluing things about women here. I put up a post with multiple examples of the latter recently. Latest instance that caught my eye was somebody's sketch of how we might use embryo selection and breeding to improve human stock: Brilliant men marry beautiful women who go along to get along, and also put out -- no, wait, misquoted the last part -- I believe the exact words were something like "have pleasant temperaments and are mentally stable."
My question is reasonable because it’s relevant to whether this group is fighting for Gaza from a place of compassion and kindness or from a place of hatred towards the other guys. This makes a big difference in how sympathetic we should be to their political demands.
Which guys are we talking about? Hypothetical people in Michigan or the people in Amsterdam who committed these crimes?
The Michigan people. Are they nice? Are they friendly to Jews in America?
My model is that most groups focus on the atrocities committed by their out-group and tend to ignore the atrocities of their in-group or far-group.
Marxists might talk at length about the genocides of the fascists and not a lot about Stalin.
Republicans talk a lot about violent immigrants and not a lot about school shootings.
Democrats talk a lot about police violence and not a lot about murders committed during the BLM riots.
The default position to take with regard to atrocities which don't fit into your groups narrative is to ignore them, sprinkled with 'thoughts and prayers' and 'let us not politicize this tragedy' to taste.
It is hard to say to what degree people intrinsically value the lives of their political or ethnic opponents, because indirect ramifications dominate. For example, I don't like Trump. However, I also think that political assassination is terrible, so I am glad that the shooter missed. Then again, if he dropped dead out of natural causes tomorrow, I would not mourn him, because I think his political effect is net negative and much larger than the value of one life. However, if I could either cause him to cease existing or being isekaied to a simulation of the White House (which is otherwise devoid of qualia) where he gets to play president, I would rather send him to that later place than condemn him to non-existence or hell.
I have one Muslim friend in Ontario who shared a story that seemed to say that the Jewish provocations before hand, e.g tearing down a Palestinian flag and hitting a taxi with a crowbar, were valid justifications to hospitalize five people.
It seems your friend believes in collective punishment
Unfortunately most people do when they're angry. Or maybe "badly-target punishment" where you punish innocent members of a group for offenses other committed.
I'm Muslim, albeit not from Michigan. My honest answer:
Probably a higher percentage than of the Jews in that crowd who would even implicitly condemn the well-documented war crimes of the IDF. In fact, they _explicitly anti_-condemned such war crimes by chanting, “Why is school out in Gaza? There are no children left there.”, literally cheering on the deaths of their opponents' babies which their kinsmen are perpetrating.
By the way, those are called fighting words, and they're not protected anywhere AFAIK. Imagine Muslims anywhere cheering on 9/11, or worse, gloating in ISIS murdering the babies of the kuffar. Would anyone be asking for the condemnation of those who attacked that crowd, even if we charitably stipulate that the questioner assumes in their question's framing without substantiating it that innocent Muslims were dragged into the violence?
Anyway, I also have a hunch it'd probably be a lower percentage than Muslims in Michigan who'd implicitly condemn any attacks on innocent Jews who didn't chant genocidal taunts, but a higher percentage than of Jews in Israel who would explicitly condemn the well-documented war crimes of the IDF.
Now I have a few questions for you. How is this on-topic? Do you find yourself questioning the condemnations of any other groups, and if so , which?
Can you explain how your on the surface unkind (accusatory), not true (disingenuous), and unnecessary (off-topic) question is actually any of those things? — and if you start by saying that it doesn't matter that it's off-topic, maybe you can also tell us about how you wanna ask the same questions of other groups in other threads, or explain how it's so necessary to ask these off-topic questions of Muslims but not other groups.
Are you fishing for condemnation? Why don't you lay out the pertinent facts as you see them when asking people about what they or their community would or wouldn't condemn, to make clear what you're expecting(?) or fishing for(?) them to condemn (no, linking to a Wikipedia page doesn't count, even if it weren't a live page)?
Finally, on the meaning of “explicit”: I've been operating under the assumption that by “explicit” condemnation you meant going out of one's way to make a statement that one condemns something before being asked, as opposed to simply being against it ideologically and/or emotionally but refusing to play the condemnation game, for example; is this what you meant?
I've also assumed that the middle point between those two positions — whereby one makes one's opposition to something clear when asked, but doesn't go out of their way to make a statement beforehand — as being implicit condemnation, too; is this what you intended?
Explicit = if asked, they'll say "That's highly disappointing" or "I condemn that" or "I hope the authorities justly prosecute those involved".
The opinions of American Jews on Gaza are not something I'm interested in, simply because I know a lot of them personally and don't need the Internet to provide me with an answer.
You've written a very long explanation of the relative degrees between Muslim condemntation of the attacks in the Netherlands and the Jewish condemnation of the attacks in Gaza. But you never once listed any numbers. What is the exact number (with error bars, if you must) of Muslims in Michigan whom if told about the events in the Netherlands will explicitly condemn them, even in a private circle of their friends? It could be 99.9%. Could be 50%. Could be 5%. I have no idea, so you tell me.
You (and others) in this thread seem very uncomfortable with the question for mysterious reasons. The default assumption is that ~99% of Muslims will condemn the attacks in the Netherlands, correct? If so, why be nervous, just say "yeah, 99% will condemn" and that's it, end of story. Ask yourself what makes this question so, so inappropriate to you.
>The opinions of American Jews on Gaza are not something I'm interested in, simply because I know a lot of them personally and don't need the Internet to provide me with an answer.
I think necessary is supposed to be about important questions that need to be asked, not questions that _you_ personally don't know the answer to, or requests for confirmation of a community's decency who you have little faith in.
Why did you even think the comments of this article were the place to get your answer? It’s a pretty specific question about one religious community’s reaction in one American state to a very recent event, you’d probably have to actually do the polling yourself.
>You've written a very long explanation of the relative degrees between Muslim condemntation of the attacks in the Netherlands and the Jewish condemnation of the attacks in Gaza. But you never once listed any numbers.
Even those relative degrees, I hedged with probably. As I stated, my honest answer. They’re my guesses based on my experience of Muslims I know and seeing things like the mockery of dead Palestinian children by Jews becoming normalised, even if not accepted.
I couldn't imagine such a large crowd of average Muslims shouting the things those Jews shouted, they'd have to be an ISIS brigade or something.
But I don’t know the Muslims in Michigan. I haven’t asked them. I don’t even know if they’d agree with you about what the facts of the event were.
The reporting on that event was enough to completely discredit the mainstream media to a community that already distrusts it when it comes to reporting on their community, especially in the current context of the ongoing genocide and the increased need for propaganda against them.
Simply put: they may not explicitly condemn it even if asked simply because they’re sick and tired of being constantly treated by the media in a way where their only two options are a) validate the media narrative, or b) look like villains.
They don’t care about the latter anymore, especially with public support being with them as the public see through the media distortions and have come to simply immediately distrust any reporting with Muslims involved where they’re painted as total villains while their ‘victims’ are completely innocent.
Of course, many journalists across various outlets have spoken about the top-down restrictions on using terms like “Palestinian”, being forced to use alternative terms like Arab instead, whereas they're allowed to say “Israelis” or “Jews” as they please, they've been exposed with the constant double standards in the use of active vs. passive voice, or infantilising teenage female soldiers of the IDF while adultifying literal toddlers (my favourite are the IDF bullets which “found their way” into a car and killed a “young lady” of toddler age), etc.
But Muslims know from daily experience how far removed the stories in the MSM are from any Muslims around them. They simply wouldn't believe the allegations of a so-called pogrom of Jews because it goes so counter to their experience of actual everyday Muslims.
So I want to stress again, a good chunk would simply not condemn it because they just have such little belief in MSM narratives when it comes to their community. And there is a widespread belief among Muslims, and I think now among non-Muslims that the MSM has a strong bias in favour of Jews and against Muslims, especially on certain topics.
They may seem to have a bias in favour of Muslims at times, but that's a bias in favour of liberals who are Muslims in name only, Muslim as an ethnicity, not as a religion.
Anyway, the point is, we know they're liars. It's no different to me believing a known and compulsive liar that I know personally and who has been exposed numerous times in brazen lies. Why should it be any different if he's got some markers of legitimacy like a PhD or a Nobel prize or whatever else, when I know him intimately and know what a fraud he is? Why look at the proxies to determine the reality when I know the reality?
I honestly genuinely couldn't give you great error bars. Higher than 5%, lower than 95%, with >75% confidence. There are just so many factors that go into condemnation beyond the content of one's beliefs.
If you want the actual belief content, every Muslim I know would be deeply opposed to attacking Jews simply for being Jews. Reports of Muhammad genociding Jews are greatly distorted by people like Douglas Murray.
I don't say this as a liberal who's Muslim in name only. I believe we should lash fornicators, stone adulterers to death, chop the hands of thieves, I'm a Muslim, and I don't pick and choose.
>You (and others) in this thread seem very uncomfortable with the question for mysterious reasons.
No one's uncomfortable, they're disgusted with you. If it's a mystery to you why that's the case.
>The default assumption is that ~99% of Muslims will condemn the attacks in the Netherlands, correct?
If that were your default assumption, why even ask? The default state of interacting with humans is that we assume the best of them until proven otherwise, without going to the extreme of unnecessarily trusting them with everything of value.
To start questioning the integrity or decency of an individual or community requires some cause for suspicion, so naturally, when you do start questioning people's character, there is an implied attack on their character.
There's no excuse for this level of cluelessness. I think you know all this.
>If so, why be nervous, just say "yeah, 99% will condemn" and that's it, end of story. Ask yourself what makes this question so, so inappropriate to you.
No one's nervous, we're not gonna give you an answer out of thin air to a very specific question regarding a recent event where the media lied (and are known to lie in general), we're not gonna validate your hidden premises (e.g. “will you condemn these attacks on Jews” obviously has an implied “innocent” there), we're not gonna accept your hiding of those premises and framing of the question in such a way that gives only the option of condemn and validate a false implied narrative or don't condemn and validate a false image of us as desiring to hurt Jews simply for being Jews or not caring about injustices perpetrated against innocent Jews.
>Ask yourself what makes this question so, so inappropriate to you.
I'm well aware of what makes it inappropriate to me. You haven't caught me out in some kneejerk moment where I have an automatic/unthinking negative emotional reaction to someone without really knowing why.
---
I'm gonna repeat a few questions you didn't answer the first time.
1. How is this on-topic?
2. Are you fishing for condemnation?
3. Why don't you lay out the pertinent facts as you see them when asking people about what they or their community would or wouldn't condemn, to make clear what you're expecting(?) or fishing for(?) them to condemn (no, linking to a Wikipedia page doesn't count, even if it weren't a live page)?
And a few new questions?
4. What is _your_ perception of the opinions of American Jews on Gaza w/ percentages of various camps?
5. What is _your_ perception of the opinions of Israeli Jews on Gaza w/ percentages of various camps?
6. What is _your_ perception of the opinions of American Jews on the spate of Muslims being attacked or killed by Zionists, Jewish and non-Jewish, including the little boy who was killed by his mom's landlord who'd always treated them well before recent events, or the Jewish woman at a swimming pool who drowned a Muslim lady's baby in the pool?
7. Which of the above incidents, or similar, were you aware of?
8. Do you condemn any of the above incidents?
9. What do you know about Islam?
10. Who taught you what you know about Islam?
11. Do you know that God doesn't command genocide in our holy book, so you have nothing to worry about when it comes to Muslims being sociologically far more adherent to their texts? This is unlike some books, such as the foundation of the oh-so-tolerant Judeo-Christian Western civilisation who were “the first” to “abolish” slavery.
1. It's a post about the political opinions of Muslims in Michigan about an ongoing war between Israel and Gaza. ~95% on topic
2. No, you're fishing for one, I'm asking a question.
3. I asked a question, links are the responsibility of those providing an answer
4. Very roughly: 90% would condemn/are upset about the civilian casualties in Gaza, 10% support them.
5. No idea, I don't live in Israel
6. If it happens in Palestine I'd say 90%/10% condemn/approve split would apply too. If the happens in the US or the Netherlands then probably 95% would condemn it, since the Palestinian question has been polarized by 10/7.
7. None
8. Of course
9, 10, 11: Islam cannot be held responsible for the actions of individual humans. People have free willl, they can't hide behind relion.
I really enjoyed this analysis; however, in this specific case I think it is missing the fact that some of these Muslims actively thought Trump was better (he has done effective outreach to them on how he will 'stop the fighting'). Now, I think they are probably wrong about that, just as the people voting for Trump because he will stop inflation are probably wrong, but that doesn't change the fact that - at least some - are voting for a candidate they see as better, not just against one they are angry with.
A journalist friend of mine spent some time with the Arab Americans in Dearborn over the election and wrote this about it:
https://www.rfi.fr/en/middle-east/20241106-trump-s-middle-east-peace-promise-wins-over-muslim-voters
I came here to say something like this. My thought (just from reading articles and stuff) is that it's not that Arab-Americans and Muslims particularly like the Democrats, but that they really hate Neocons. Dick Cheney endorsed Harris, not Trump. It wouldn't be surprising at all if just that endorsement repelled a significant percentage of Arab-Americans away from the Democrats. You don't need complex game theory: one of their worst enemies endorsed Harris.
https://www.aljazeera.com/opinions/2024/9/12/cheneys-policies-as-vp-caused-immense-human-suffering-on-a-global-scale
(TLDR: Trump is bad, but Cheney is world-historical, genocidal evil bad)
Is Cheney Machiavellian enough to anticipate this reaction and endorse Harris in order to boost Trump's chances?
Does he actually like Trump enough?
> (TLDR: Trump is bad, but Cheney is world-historical, genocidal evil bad)
I find myself roughly agreeing with the AJ article that the Bush era war on terror was a catastrophe on many fronts. The long term outcome of either the Iraqi or the Afghanistan invasion were bad. Gitmo was terrible.
However, when I think world-historical, genocidal evil my yardstick is Hitler, and GWB and his cronies are just not playing in the same league, evil-wise.
Gitmo is evil, but it is not Auschwitz in either scale or intensity. GWB has earned his place on the list of post-WW2 leaders causing large scale human suffering, but without having checked the numbers, my gut feeling is that Vietnam was likely worse.
In particular, I think that the framing of Cheney as genocidal is inflammatory and thus should be backed up by evidence. Even though AJ is not shy about making inflammatory allegations of genocide (e.g. against the IDF), they do not make such allegations against Cheney.
Yes, I think this probably explains more of what happened than some complicated game theory take. I haven't seen a ton of exit polls specifically breaking out Jewish/Muslim voters, but Trump winning Dearborn seems pretty surprising to me. It seems that the Biden/Harris administration got themselves blamed by all sides (perhaps unfairly); but at the same time, it's not surprising that voters would say that Trump's 1st term saw a much more peaceful Middle East than Biden's term, and for them to vote accordingly.
What we didn't see was any sort of decisive break to the Green Party, or any other third party choice. I don't think Trump won any states where a combo of Democratic+Green votes would've beaten him. It seems to me that voting third party would be a better way to signal your power as a voting bloc: if Trump's votes go up, you might just assume he turned out his base better, but if an interest group comes out and says "We're voting for Stein" and then she does way better than the Greens normally do, you can more easily chalk up the difference to that one interest group.
I think this is really it. An anti-incumbency vote against the administration that perceived to be prosecuting the conflicts with a sprinkling of cultural topics contributing. Harris failed to distance herself from Biden so they voted for the other guy.
And religious people tend to be conservative + conservative people tend to vote for Republicans. Before 9/11 and the war on terror Muslims leaned towards Republicans.
I read a twitter take that was something like this:
For years there's been jockeying for status among the Democrats' intersectional coalition. But recently two clear winners have emerged: black women, and transsexuals, with all other groups demoted to a subsidiary role within the coalition. As a result a bunch of other groups (this particular tweet was about Latino men) are abandoning the Democrats.
> And religious people tend to be conservative [...]
Conservative in what sense? The current Pope is rather religious [citation needed], and he sounds like a commie (in line with what most of the Bible says).
I don't particularly get formal decision theory, but similarly feel somewhat flummoxed and disappointed at the perennial non-catered-to-voter's choice of "settle for whatever slice of loaf coalitional politics is willing to toss my way" or Threaten To Vote For Someone Else. It can definitely lead to some deeply uncomfortable places to be in that taken-for-granite bucket, being expected to go along with outrageous nonsense you wholeheartedly disagree with because You Should See The Other Guy / No Candidate Is Perfect (said in patronizing voice by a Very Serious Adult who compromises with sin). And, yeah, in places like CA where the opposition party is mostly a formality, there *isn't* always a Someone Else. Or they're so laughably below A Literal Heuristic Rock that one is inclined to simply not vote (but then that weakens the ostensible feedback mechanism). Organization helps, to some extent, but...that just kicks the coalitional politics can down the road. Building a smaller tent out of those who self-select or are rejected from the bigger one is gonna have some gnarly selection effects by default. No orthodoxy for the heterodox!
(Vote-swapping compacts seem like a clever solution to this dilemma, but pose other large coordination difficulties that I think in practice make them impossible to scale effectively. The mercenary attitude necessary for being willing to bargain one's vote necessarily makes counterparty trust a bit tricky.)
Voting isn't really the main vehicle by which most interest groups get what they want. Small but organized groups can extract benefits for themselves at a cost to the overwhelming majority, usually through means like campaign donations, lobbying, protests, getting their members to call or write to politicians, getting their members appointed to boards and departments, using their expertise to suggest regulations or laws, etc. Also, in many cases they can do these things regardless of which party is in power. The ability to deliver votes is usually minor in comparison.
Yes, (individual) voting seldom changes much.
> So I think this is the right move if everyone involved is a superintelligence and has already made all commitments they would make if they were omniscient and had infinite time to think about the problem.
I don't think it does work, because exactly the same logic applies in reverse to voting for Trump. Trump's offer is even lower, so they should punish him by voting for Kamala with even higher probability than the reverse. And in real life, if you ask the leftists who keep trying this strategy why they don't try to influence Trump instead, they say it's because they think Trump can't be influenced to support them, and the Democrats can, which seriously undermines the idea that this isn't threat-like behaviour.
(The Palestine-supporters are by no means the first leftist group to try this, I've been hearing leftists talk about how they should punish the Democrats for not being leftist enough since at least 2016, and the Palestine-supporters almost certainly got the idea from there)
Maybe that's why Eliezer recommended abstaining or canceling our votes instead of voting for Trump, which maybe wouldn't have this problem? I admit I'm very confused about it though.
I think Eliezer's notion of "threat" relies heavily on having a privileged null action. If you pick "don't vote" as the null action there's no longer a contradiction, you just don't vote for anyone, but I'm not convinced that really makes sense either.
You must be young. The entire Ralph Nader 2000 campaign was about punishing democrats for not being left enough.
No it was not and it has never explicitly been that until maybe this very election, and just barely. That's what it "should" have been about perhaps, but instead it was mostly a completely moralistic argument that you shouldn't have to vote for someone with whom you disagree on almost everything, combined with some vague hope that a third party could come to power, which very few take seriously. I've never heard a leftist actually articulate an argument to vote the opposite way until just a few weeks ago in this interview with Noura Erkat: https://youtu.be/bqQ75cktTY8
"I want Kamala to lose". Still it was very rare to hear people expressing stuff like this publicly and many refused to vote, voted third party, or voted for her anyway. What's different about this time? Everything. Kamala made 0 concessions to anyone left of center and in particular to those against the Gaza war, none, when asked how her administration would be different from Biden's she answered that she couldn't think of anything.
Really, to a leftist it's 99% to 100% Satan, not just a cute thought experiment, there's a utilitarian argument to be made that settling for that extra 1% is worth it to get someone who's double digit not Satan in 4 years, and to hold them accountable with the threat of voting Satan if they try to sidle up to him again.
I personally saw leftists in 2016 insisting there would be no reason for them to vote if Bernie Sanders wasn't the candidate, or if the Democrats didn't adopt a socialist platform, and making similar demands in 2020. AFAICT nothing has changed here except the specific demands. This is particularly notable because Biden in fact made substantial concessions to the Left by e.g. passing a huge economic stimulus, dedicating a bunch of money to renewable energy, trying to forgive a huge amount of student loans only to be blocked by the (republican) courts, and to my observations this didn't reduce the amount of Leftist demands at all. I even recall people explicitly saying they would respond to policy wins by increasing their demands.
It seems to me that the people making these demands have, at best, drastically overestimated the size of their voting block, and more likely are just more committed to railing against the democrats then they are to accomplishing any policy outcomes, and either way there is no reason to make any concessions to them at all.
Oh they absolutely did talk about not voting. In the end though it was the same argument, they couldn't bring themselves to vote Biden because they recognized he hated their guts and wouldn't do anything substantial despite some token rhetorical concessions. This is actually what happened btw if you were paying attention... And anyway a lot of these people, probably most, would later admit to voting Biden anyway.
Regardless, I'm talking about a strategy of voting the other side, assuming you're in a swing state, as in you are explicitly voting for the outcome you desire which is for the person you hate even more to win. You would never have heard this mentioned in any kind of even remotely public left discourse in 2020, it just barely started happening this election and only right at the very end.
Yes, 2016 is also the first presidential election I was following Leftist Thought for, so I have no reason to believe it doesn't go back much further
There are always leftists who say "we won't vote for the Dem because X" and X changes from student loans, to universal health care, to Palestine, and even if that was solved there would be a brand new thing.
The centrists have a much better chance of bluffing. They are, by definition, less bothered by a Trump victory. Biden's position was closer to what they wanted than Trump, but even Trump's position isn't unpalatable.
>“here are the usual levers for influencing Democratic Party policy, like voting in the primary, you lost, and you are trying to re-litigate a completed political process because you don’t like the result, in the same way we condemn Donald Trump for doing.”
This argument would have been more effective if there was a competitive primary this year. (for the record, I voted for Dean Phillips, but the best potential Democratic candidates didn't run out of deference to Biden)
I am very confused why this is framed as decision theory rather than game theory. If you view it as a repeated game, standard game theory tools make sense here. In particular, accepting one cent isn't necessarily a Nash equilibrium of the ultimatum game if the game is repeated; you make yourself worse off now to get better offers in the future. Muslims are doing the same thing this election in this example; they're making themselves worse off now in the hopes of getting better offers in future elections. I really don't see why you need to throw on so many epicycles to explain it.
Exactly! I was thinking the same thing. Eliezer's answer seemed overcomplicated to me (also I really dislike it when people like using so many technical terms that it obscures a lot of the thought process). Yes, in artificial experiments where you know the game doesn't get repeated it is rational to accept 0.01. But this almost never happens in reality and even if a particular game doesn't get repeated, you will play other games and as long as your behavior can be observed of any party you might play new games with, it is still better to refuse to maintain a reputation of someone who cannot be mugged in a game like this. And perhaps more importantly, this is mostly instinctual behavior in most people most of the time (and not just people, some smarter animals have this hardwired too) and it is not worth thinking to hard in an experiment where the total prize is negligible. I would expect different results if this is a prize of 1 billion and you're offered 1 million or nothing (also because a game with those stakes is unlikely to repeat and it is so obvious that you act accordingly).
I also agree that Yudkowsky’s answers are low signal to noise ratio and these seems like a pattern with him. Of course this is a repeated game, and defecting makes perfect sense as long as your gripes with the Republicans are not existential.
Actually, as someone who was in that massive Discord thread, I think Eliezer's answers here are still less specified than I want. This stuff is tricky, and there's a lot of moving parts, and I only understood some and have forgotten much of that.
Like, one of the things with your $999m to $1m split is, how aligned are you with the other player? A little quote that I'm proud of: "the greatest trick the devil ever pulled was convincing the world that non-iterated games exist". To a close approximation, you'll always have another interaction, and how confident are you that the other player won't use that $999m to act against your interests in a way that you can't resist with a measly $1m?
> Eliezer's answer seemed overcomplicated to me (also I really dislike it when people like using so many technical terms that it obscures a lot of the thought process).
I suspect that's because he assumed a lot of shared context that you don't have?
No, it’s just really not a good answer. Lots of jargon and epicycles in his examples, but not very illuminating versus simply applying *really* basic game theory. Even Scott doesn’t seem to find it very satisfying (“I think that’s what Eliezer means…”) though I think for personal relationship reasons he extends Yudkowsky a lot of credit that he has not earned in this realm.
I agree that in this case, modeling it as an iterative game is fine, but in general it's still worth understanding why it's good to cooperate with the stranger you meet only once and never again.
I think one of the points of the Yudkowsky style decision theory is to somehow merge decision theory and game theory. If we treat all the different players as somehow algorithmically correlated, we then just have to optimize that algorithm against the background of whatever is left, rather than treating the different players as separate.
I don't think the war in Gaza is a genocide (though it might be some sub-genocidal level of bad, IDK), but if you did think that and you identified with the victims, it would probably be more important to stop/mitigate that in the next four years than get whatever unknown and hypothetical concessions you think you might be able to get after the genocide is completed in the next election.
First of all, yes, in general, it depends on the relative values of what you're losing now and what you might gain in the future. It could be the case that preventing harm now is more important than future gains, but it could also not be. In particular, it could be the case that the belief is Trump will make the war worse, but it's going to be bad either way, and the national impact of Trump is small. In that case, building the reputation and making future gains could easily outweigh the present harm.
>In particular, it could be the case that the belief is Trump will make the war worse, but it's going to be bad either way, and the national impact of Trump is small.
I agree with this, and that this would go a long way to explaining the phenomenon in question. But for the record, it seems like a just false empirical belief to me, and most of the people accusing these voters of being irrational (rightly or wrongly) are trying to point out to them that the war will be much worse for Gazans than it already is under Trump. I guess we'll see.
Yeah, in a sense this is the entire reason behind war? People are willing to potentially sacrifice their entire future to prevent something bad enough from happening right now.
In the repeated ultimatum game, played by two CDT agents, the last round plays like the single-shot version. Since they already know how the last round plays, they know anything that happens in the second last round has no impact on it, so it also plays like the single-shot version. By induction, every round has the offerer offer the minimum value, and the decider accepts it.
This is essentially the same reason why CDT agents mutually defect on every round in the repeated prisoner's dilemma.
To get some other outcome, you need some other theory of decision making, that cares about things other than just the causal effects of its actions.
Those are some staggeringly dumb agents
"CDT is a bad decision theory" is a perfectly sensible response to this, just, you then have the problem of picking a better decision theory.
I hate to spoil the nerd party, but has it occurred to anyone that maybe the Muslims are genuinely just mad and that there isn't a game theoretical reason for defecting?
( anger itself is somewhat game theoretical, but the incentives acting on an individual are about how they think other people will view them based on what they've heard on social media, which is Game Theory but with such skewed and distorted inputs that it's really more of Sociology than strategy.)
Not to mention the fact--easily ignored on a place like this but highly salient--that most Muslim Americans motivated by that identity are varying levels of highly religious, and therefore believe that their deity is directing matters toward an ultimate end whose shape and characteristics they already know (Judgement Day).
On a related note, most religious Muslims are highly conservative, and only support the Democratic coalition because they expect the Republicans to oppose Muslim religious freedom and immigration. In terms of other social issues, they tend to be as opposed to Democratic policies as conservative Republicans are. Democrats are very much an ally of convenience for them.
As you say, evolution obviously programmed us to feel anger for game theoretic reasons. And I think it's worthwhile to understand why evolution did that and maybe spot places where we shouldn't indulge in our anger. Maybe the Michigan Muslims don't care about this question, but *I* care about knowing whether their strategy is correct from a game theoretic perspective.
I follow one of the people pushing this movement, and it’s pretty clear he’s approaching it from a game theoretic perspective, even if not formalising it in such terms to his audience AFAIR.
It’s not just a phenomenon of American Muslims lashing out due to anger, they’re thinking about the issue from a rational leverage perspective.
I don't especially care about what's happening in real life, I care about the solution to the problem they've unwittingly raised.
This is the most rationalist sentence I've seen you post in some time, but I feel like if you don't qualify it it'll definitely show up in the next newspaper profile of you.
Hopefully we're all friends here, either that or fully-aligned journalists.
How’s it unwitting?
Fair point. I would throw a similar idea out there: I work with a Palestinian guy (n=1 but I don't think he's all that unusual). He is, as you might expect, highly socially conservative. You can imagine someone like him being willing to cast his lot with the Democratic Party despite his deeeeeeep disagreements with them on most issues if he feels it might be of some benefit to his family back home. Take that away, though, and he's basically a Moral Majority devotee. American Muslims: not single issue voters!
If you immediately know the candlelight is fire, the meal was cooked a long time ago.
In other words, a solution that avoids this problem is, "Start organizing and advocating in useful directions in 2004-2016, when it was already becoming clear which direction things were moving and that the democrats lacked a deep bench of candidates that were actually able to competently represent their constituents." Then, in the cases where it happens that you still get a candidate you can't stand, you can either support them anyway without such long-term decision theoretic implications, or vote against them and trust that there are enough other strong democratic candidates in other offices that you're not throwing the country to (what you may see as) the wolves.
Even assuming a time machine, I'm not seeing a concrete actionable plan here
We look around us right now, at trends in both major parties, and start trying to influence (or create) politicians in both parties who think about things in similar ways.
The Ultimatum Game comparison seems flawed because it's a positive-sum game. Both participants can win money if they find an agreement, or nothing.
The voting situation regarding Israel seems negative sum to me. Harris can either appeal to pro-Palestinian voters, and lose pro-Israel votes, or appeal to pro-Israel voters and lose pro-Palestinian votes. Assuming that there are more pro-Israel than pro-Palestinian voters for Harris to lose, then (a purely cynical version of) Harris isn't gonna do anything for pro-Palestinians no matter how hard they threaten to vote for Trump, and voting for Trump is a pure waste that doesn't give anyone incentive to help them.
I think this is a common failure more of leftist ideology when confronted with centrist blocs. Leftists often show contempt for centrism/pacifism because they assume the only possible option a reasonable person can have is "do nothing" or "side with me", and they resent people for picking option 1.
That doesn't make the situation negative-sum for Harris, just one with tradeoffs. Harris can still set a policy on Israel that she expects to get her as many voters as possible. Voting blocks can try to shift where Harris falls by partially or probabilistically withholding their vote unless Harris crosses some particular threshold.
(If the voters are naive like CDT and think they have to vote for "whichever electable candidate has the better Israel policy", and don't consider "don't vote / vote 3rd-party" as an option, then Harris can collect all of their votes via the tiniest possible policy shift away from Trump.)
It's true that that doesn't necessarily make the situation negative-sum.
But then again, politicians often feel that taking a position is a negative-sum thing to do - that any position they take will get them less support than a positionless feel-good platform - and I suspect they are usually right about that.
There's a scene from the sitcom Friends that's stuck with me:
In this episode, the audience learns that Phoebe is secretly married to a Canadian ice dancer, Duncan, for purposes of making it easier for him to live and work in the United States. The marriage is one of convenience; he proposed to her as a gay man who would have no relationship with her other than the existing friendship and a formal state of marriage.
But, we learn, Phoebe agreed because she was secretly in love with him. The reason we learn this is that Duncan has shown up to ask Phoebe for a divorce - he's realized that he is straight, and he wants to marry a different woman.
Phoebe is naturally upset. And when she agrees to the divorce, she can't stop herself from asking "if you had realized you were straight earlier, do you think I would have been the one..."
At which point she interrupts herself, saying "you know what, I don't think either answer would make me feel better."
This seems to be a plausible example of a situation where knowing more about it always leaves you worse off, regardless of what it is that you've learned.
> The Ultimatum Game comparison seems flawed because it's a positive-sum game. Both participants can win money if they find an agreement, or nothing.
We typical compare fixed sum games and variable sum games.
We typically call the former 'zero sum' games, but any fixed sum yields the same analysis.
For variable sum games, applying any common additive offset to all outcomes doesn't change anything about decisions. (That's like: 'you find a coin on the pavement before the game begins.')
Scott, what’s happened to you? You’re ignoring all the reasons Muslims might actually _like_ Trump and adopting, without any questioning or criticism, the brain dead narratives of the New York Times crew.
The man expanded his coalition among basically all groups, because he actually tried to do this. The Harris campaign’s whole argument was, “orange man bad”, and Trump said he was going to end the conflict quickly. Doesn’t matter if you think the guy doesn’t like your religion - what matters is he’s at least saying the killing has to stop and you suspect he might actually deliver.
Rather than assume I'm trusting any "narrative", you can just listen to what these people say. https://abandonharris.com/faqs/ , second video, "What Happens If Trump Wins".
I understand people are getting overly twitchy because of the election result, but I request an apology for your first paragraph.
So there's one guy who has a game-theoretic argument, which may or may not be sincere. And it looks like he's persuaded at least 25 other people to go to a rally at some point. But you keep generalising from this one guy to the voting patterns of hundreds of thousands of other people, you talk about "Michigan Muslims" instead of "Hassan Abdel Salam and anyone who happens to be listening to him".
Besides, the whole game theory analysis only makes sense if you assume that Michigan Muslims would "naturally" vote for Democrats. Muslim Americans tend to be socially conservative, working class, religious, and legal immigrants, who would tend to be more natural Trump voters. On the other hand you've got Trump's "Muslim ban" policy of eight years ago which would tend to push them away from him, but that's not a policy this time around. So it's fair to model Muslim voters as people whose preferences are not particularly well aligned with either party.
I think the spokesperson for the main organization representing this group is a credible source for their views.
Muslims in the US are destined to politically walk forever through the valley of the shadow of death, I am afraid. To threaten that “if you do not give me something more, I’ll vote for those who will give me even less” is not a credible threat. No amount of over-thinking can change that.
Maybe they can get something by splitting their vote, though. Something like “unless you give me something more, I’ll only vote for you in presidential elections - in Senate and House of Representative races, I will vote for you opponent. “ Assuming that securing the President is more important for Muslims (Presidents have a lot of leverage in foreign policy), while winning in Senate and House races are less important for them (since that's more domestic policy stuff), but quite important for the Democrats.
It’s not likely to be effective, either, but at least it is something.
The voting dynamics are interesting to think about, but in the real world there are a lot of other payoffs. They're are political appointments, policies, pork, election resources for future races, and probably many others.
If you are helpful in the coalition, then you can jockey for those other prizes. And in the case of the Michigan Muslims, I think they probably had some hope of such a prize from Harris, even if forcing Israel to give up wasn't in the cards, and no hope at all from Trump.
The rage toxo post (aka best SSC post ever) suggested that we hate apostates more than infidels. I think that's what's going on here. They felt that Harris should have been on their side but wasn't, so were way more mad at her than at Trump, who is ideologically far away and therefore there's no sense of betrayal.
> pork
Well maybe not pork
I feel like I can now retire from the Internet now that Scott (and Eliezer!) have responded to an offhand comment. https://www.astralcodexten.com/p/acx-endorses-harris-oliver-or-stein/comment/74832519
My only incremental contribution is to agree with Scott insofar as the fact that this is not an obviously solved question makes it possible that almost anything is a "strategy." This is just a variant of the Folk Theorem of game theory, that almost anything can be a dynamic equilibrium with sufficient patience.
This seems overly complicated. For all the overheated rhetoric, few people actually think that this is going to be the last election. As such, it’s just iterated game theory, and the game being iterated is a sort of ultimatum game. If you think you were given a shitty offer you reject it and eat the short term cost in the hope that in the next round you’ll be offered a better deal.
There’s all that needs be said.
> For all the overheated rhetoric, few people actually think that this is going to be the last election.
It's probably the last election for Harris, and hopefully the last election for Trump.
I'd been thinking about this too (not with Israel, but with policy in general) and ran into two big questions:
1) Shapely Values are notoriously non associative. If I calculate Harris and me, it's reasonable that I get nothing. If I calculate Harris and People Like Me it's more complicated.
2) To what extent should Trump be seen as a threat by the parties and ignored? In a normal election I'd say yes: the shadowy elites that run both parties, while not literally one person, use mental models of each other to act as a single unit and maintain strict solidarity. But Trump is, maybe, not part of that coalition. Though the coalition still decided to back him. So I don't know.
I would think the best approach is to try to coordinate some sort of write-in campaign. The Prophet Muhammad, Yasser Arafat...doesn't really matter who; you're just trying to send a message to both parties that there were votes up for grabs that went to nobody.
This also works.
Part of it is punishment, sure. I mean that seemed to be the collective strategy: You have taken us for granted and not served our interests, so we are going to punish you.
But an equal or larger part is: You have shown you cannot solve this problem, so we are going to take a (maybe foolish) chance that this other guy can.
> One of the few checks that voters have on the Democrats’ corruption and incompetence levels is to threaten to vote Republican. Is there some level of Democratic corruption at which I should vote Republican to “punish” the incumbent even if I think the Republican would be a worse leader overall? I’m not sure, and currently lean towards no, but I can’t say it doesn’t tempt me.
I really think you're suffering from failing to distinguish trade offers and threats. There are levels of Democratic noncooperation where you vote third party for a candidate who offered you more -- that is, throw away your vote, in the present system, but in an organized way that shows your power to show up at the polls and your responsiveness to offers. It never makes sense to vote for Trump (unless he's made you a better offer). I would expect actual political operatives to treat these bargaining positions differently, just like a small-town market outside of California differently considers the positions "I'm not shopping here if you offer me too few grapes per dollar" and "Gimme more grapes or I'll break a window."
> I really think you're suffering from failing to distinguish trade offers and threats.
I think what you mean by 'threat' is sufficiently different from ordinary senses that it might be useful to invent or repurpose another term to more clearly differentiate what you mean. (One of) the dictionary definition(s) seems to cover what you mean but ordinary usage has expanded far beyond its narrowness. Sadly, I couldn't think of any good and pithy alternatives.
Some not-particularly-satisfying alternatives/qualifications:
- extortionate threat
- punitive threat
- conditional threat
Or just something like "LDT-threat". Unsuited to casual conversation, but that's fine.
Maybe – but it seems perfectly understandable and relevant to other decision theories and game theory too. A non-extortionate threat should be (potentially) simply avoided or its negative consequences mitigated, whereas an extortionate threat should also probably be disincentivized too.
> Trump is left actively better off by his decision to be so hostile to the Muslims
I don't think this is a correct. If Trump were less hostile to Muslims, presumably they would be even more willing to vote for him. So there is no incentive for him to be hostile.
I think a better model here is that one candidate has policies A, C and another B, C. The voters prefer "A and not C", and since "not C" for them is much more important then A, then they flip to the second candidate.
If the voters in question actually liked Trump's other policies, then following the same logic they
should have switched to Kamala to punish Trump for an incorrect stance on the Muslims.
If Trump was less hostile to Muslims, Muslims would threaten him saying that Trump should be even more pro-Muslim or else they will vote for Harris.
But because Trump is hostile to Muslims, it doesn't occur to anyone to make that threat. It would be pointless.
But he's not hostile to Muslims this time around. Eight years ago he was promising to ban visas for all Muslims. This time around he hasn't talked about it. Part of that is a lack of recent Muslim terrorist attacks in the US, but part of it is that everyone knows Muslim votes are up for grabs in Michigan.
We could also look at the other side of the game theory here -- does the threat of Trump's threatened Muslim ban dissuade terrorist attacks in the US?
"Well he hasn't threatened me TODAY, so why should I have a problem with him?"
If you didn't protest-vote in this election, given everything including the Democrats and most of your personal media lying to you for at least months and reasonably likely years about Biden's mental faculties leading to a candidate who nobody wanted in the first place, I don't think you're willing to protest-vote at all.
So, if you acknowledge a protest-vote is valid, I think your first step is noticing that you need to be a lot more willing to protest-vote in the first place.
More generally - if you're a member of a coalition, and whatever interest group you believe yourself to represent isn't doing better as a part of that coalition than if you weren't part of that coalition, that's a good time to protest vote; your interests are being disregarded.
As a more useful metric: If your voting group is being pushed to go out and vote, that is the time to protest vote. If your specific group is being specifically pushed to go out and vote, it's because the party apparatus knows that your group isn't very enthusiastic about the offer being made and probably won't show up to vote in their usual numbers, which probably means that your group is being offered a raw deal by the coalition.
So, for example, the Democrats were really pushing men to go out and vote. Well, the party is still operating as if it is the early 2010s and it's socially acceptable to shit on men while demanding they support women (Christ, I haven't seen so much use of the phrase "toxic masculinity" in nearly a decade). Unsurprisingly, men turned up and voted against them, because by and large we're done with this bullshit. I'd hazard a guess that the misandry inherent in the election was one of the key mistakes of the effective campaign.
Interesting post. I had a similar thought a few days ago ( https://www.griffinknight.com/p/voting-to-send-a-message ). I don't really go into the game theory of it, but I tried to think through some sort of framework for when "voting to send a message" is a viable strategy in that it makes things "better" for you (of course what is better is subjective based on the voter).
I think the right questions to ask oneself when protest voting are: 1) is either outcome of the vote relatively unimportant and non-consequential to me? And 2) am I confident that my intended message will actually be received and actionable? If the answer to both of these is not an astounding YES, it probably just makes sense to vote for whichever direct outcome you think will be "better" for you.
The limitation of this kind of analysis is that it treats politics as a constrained optimisation problem in a simple closed world. But strategy does not exist in a closed world. Finding a good strategy is a problem requiring creativity. Even in something as simple as the micro mouse racing competition, creativity dominates optimisation.
The real world is overwhelmingly complex. A good strategy comes first from a diagnosis that reduces this complexity, and a guiding policy that governs how you act. If you know everyone's reduction, perhaps this is simple enough to do constained optimization.
It's true that Democrats reduction was leading them to ignore the Muslim vote, and that trump is worse. But that meant Muslims had one choice: try to get democrats to think again, to move outside the current paradigm. They failed, but that doesn't mean it was wrong to attempt.
The Democratic policy was based around the idea that they should avoid responsibility for Gaza. Give Israel all the weapons to satisfy the Israeli lobby, make gestures towards humanitarian concerns for the Muslim vote. Don't attempt to look for other solutions, lest we get the blame.
But this ignores that the Israeli governments actions are bad in the long term for Israel, the US, and the Jewish diaspora; and the Israeli public support it largely because they have lost hope. Any strategy that stops the ethnic cleansing in Gaza must involve convincing these constituencies that a better way is possible, but that won't happen if all you are prepared to offer is weapons and hand-wringing.
I think that most of these analyses don’t adequately consider the long-term consequences of the relationship that is being formed …
In a martial contest, if person A does not in some way threaten person B, then B will eventually wear down A with a series of attacks and A will always lose. Even when defending, a person must show the threat of a counter-attack or they are doomed. There is a 2500 year history behind this theory, e.g., Sun Tzu.
So unless the Michigan Muslims want to trust in Harris’s commitment to “do the right thing”, they must offer a credible threat. If Harris does not in some way fear the Michigan Muslims then she will continually give away pieces of “their pie” and they will be left with a total loss. Since “do nothing” creates a long term certainty of total loss, and absent any other potential courses of action, their action seems well chosen.
Once the opponents are balanced then negotiations can start.
Not going to comment on the Muslim issue specifically, but with some factions on the left, there’s a real problem of them taking any compromise and using that as the basis to demand more concessions. After a few rounds of that, those factions lose all influence.
I've heard that Brexit happened that way. That a solid majority was against Brexit, but many of them decided to vote for it anyway to signal disagreement with David Camerons policies.
I don't know whether it's accurate. There were certainly protest voters like that, but I have not investigated whether it was really unusually many .
I don't think that was really true. What did happen is that there was nothing that forced the Brexit advocates to define their offer in concrete terms, so they could portray it as the answer to whatever you were irritated about -all things to all people. So they appealed to both people who wanted protection from the international markets and people who wanted less. Social conservative protectionists and free-traders.
Remain had a narrow 52-48 lead in the polls, which turned into a 48-52 loss when votes were actually counted. This is easy to explain with ordinary systematic polling errors (shy Tories and the like).
In a real election, I'm skeptical of any of the probabilistic stratagems. Even if they are perfectly executed, the candidates can never be sure that they actually _were_ executed.
Hmm... Re the Electoral College, perhaps it is time for a coalition of disempowered non-swing-state voters to organize for a popular vote...
An alternative is a coordinated vote for someone else, a third-party candidate or fictional character, especially when the person to vote for is arbitrarily decided in public shortly before the election. Just like protest marches, they convey the unspoken message "I can mobilize a whole lot of people who do what I want them to do, therefore I have intent and means, and therefore I am someone worth negotiating with".
For fun, take a look at this webcomic; page 87 is where I get the "intent and means" language from.
https://well-of-souls.com/outsider/outsider085.html
Many Thanks!
>An alternative is a coordinated vote for someone else, a third-party candidate or fictional character, especially when the person to vote for is arbitrarily decided in public shortly before the election.
Yes, that would convey power, but also note that it is a deterministic action. For any of the _probabilistic_ strategies to be credible, not only would the leader(s) have to coordinate the vote, they would have to publicly declare the strategy, publicly show the resolution of the random part, and this is on top of the actions you describe. I think it would be too hard to do. The deterministic part, by itself, is already hard to do.
Yeah, probabilistic is not so good when performing for a human audience. People like feeling as though their leaders have a plan, and we don't all have IQs of 130.
Many Thanks, Agreed!
Talking about taboos and savvy brings us out of game theory, but since you brought it up...if your block never breaks taboos but just relies on being savvy and accepting handouts proportional to power, you risk being rolled by people who actually are savvy and trick you into believing you don't deserve much.
Yes, it's dangerous to deal with entities that have greater intelligence and/or greater social ability. That's kinda where all this came from. ;-)
The correct strategy is obvious. Vote for Harris ('cuz she ain't Trump) but claim you are voting for Trump, and can be wooed back.
Hard to coordinate, though.
Is this what happened to the black vote? It used to be reliably Republican, then some of them started saying "actually the Republicans aren't satisfying our interests enough, we're going to vote for the Democrats despite the slavery and Jim Crow and shit, wink wink". But enough of them didn't hear the "wink wink" that they went and became a reliable Democrat voting bloc, whose interests still aren't being satisfied.
I think it was LBJ pushing through the Civil Rights legislation that moved blacks toward the Ds. And then the local white backlash was embraced by Rs, which further moved blacks toward the Ds.
One problem is that vote totals are public. So politicians will look back and see "not many folks here voted for Trump", and assume that you're irrelevant.
Eliezer Yudkowsky is not actually an authority on game theory. His interpretation that claims Logical Decision Theory isn't just bunk is not supported by real experts. I like a lot of his work but please don't treat him as a good game theory consultant instead of an actual economics professor.
To he clear, Eliezer is a published reseaecher in this branch of philosophy/mathematics.
More to the point, if his arguments are wrong, it should be possible to articulate *why*. (And then once you do that, the argument from authority becomes irrelevant to anyone with the time, patience, and intelligence to judge the debate!)
I can articulate why his arguments are wrong.
1) Like he says in his arguments, he has not actually mathematically proven his research, "First of all, because nobody has actually derived the LDT Ultimatum solution from first principles.", he's just going off vibes that he thinks his conclusions should be true
2) If LDT works, then you should be able to set up simulations/experiments where people/bots behave according to LDT principles. You can create all sorts of experiments with lots of bots operating under different strategies play Prisoner's Dillemas or other games, and the winners are consistently what CDT predicts they'd be. When humans play, sometimes results differe, e.g humans rejecting "unfair" offers in the ultimatum game, but I think that's simply explained by humans feeling like reputation matters even when in the rules of that specific game it doesn't.
3) I think the fact that mainstream economics doesn't agree with him, on top of the fact he admits he hasn't rigorously proven LDT, plus the fact there's no experimentalevidence supporting it, adds up to a very strong case that it's bunk.
It's certainly an unusual strategy, which means that either they're missing something or every other interest group is (or there's something else unusual going on in this specific case, but I have no idea what that would be). "Not voting at all" seems like something that other interest groups might do and feels a lot less extreme, even though it's not really that dissimilar from half as many people voting for the opposition instead.
Maybe they're too new of a group to have other forms of influence? In my model of, say, a teacher's union, if they're unhappy with the local Dem politician, they might fund a primary opponent, organize protests or campaigns to call/write their office, write letters to the local newspaper, threaten a strike, etc. I don't know if they thought these were not viable options. Overall though, it seems like the strongest incentive is the one that says "we don't negotiate with... defectors" and just ignores them.
I think there might be too much mathing going on here and not enough thought about human nature around personal wants and tradeoffs.
Has anyone ever run a version of the Ultimatum Game with an actually important amount of *real* money on offer? Because I think the official studies made a big mistake in offering trivial amounts of cash.
Ten or fifty or even a hundred dollars has almost no value to most people, and thus if the percentage offered is insultingly small, person B might very happily consider "losing" $0.01 or $2.50 or $10 a good value to punish the other person for their "greed."
But if the amount of actual real money on offer is $100,000, "unfair" split offers of 90/10 or even 99/1 are going to be a lot more tempting for person B!
I know I'm not willing to forgo $1000 merely to prevent someone else from having $99,000. And I'd be *delighted* to have $10,000 free dollars for doing nothing except saying, "sure, keep $90k."
If the amount of real money on offer is $1,000,000, I'm willing to take an even smaller "insulting" percentage.
I have a strong intuition that the distinction *REALLY* matters. The Ultimatum Game doesn't show us anything except that people will cheerfully spend trivial amounts of money to punish others because they see it as a good value.
This may indicate that the Muslims didn't think Harris or the Democrats ever had anything "real" to offer them, and that they might as well punish them for not offering / delivering more.
I asked pretty much the same thing you did in a graduate seminar I took almost 20 years ago, with the same thinking behind it. IIRC, the professor's answer was that there are studies done with medium sized prizes (on the order of hundreds of dollars), with test subjects living in poor countries for whom the prize was quite significant relative to their incomes. In those studies, people were still willing to reject very unequal divisions, but were more apt to accept moderately unequal divisions than the same pool of subjects were with smaller prize pools.
Huh. Would those studies have been able to control for more reputation/honor-based cultures, which many poor countries are likely to have? So that there was a greater perceived value in saying "fuck you" to an insultingly low offer?
But more importantly, was the prize amount in the poor countries actually *REALLY* significant, like, buy-a-car or launch-your-kid-to-a-better-life or never-go-hungry-again significant?
$1000 is a lot of money to me, but I still *might* say "fuck you" to someone offering it out of a million.
But if I was offered $100,000 out of $10,000,000 to pay off my mortgage, I would certainly take that deal. I would take $100,000 no matter *what* the deal was. There's no amount of "fuck you" satisfaction that would be more satisfying than having $100,000.
It's hard to imagine other people not agreeing when the money is really, *really,* real, but it's certainly not the first time I've been wildly wrong modeling other people!
That, however, would rule out human nature as a linear system of equations, which would make it much more inconvenient for academic researchers to "solve". So we can't have that.
Also, good luck getting a large enough research grant to allow doing million-dollar ultimatum games enough times to draw statistically useful conclusions.
I think the controls were running the game with different prize pools amounts with participants from the same culture, which seems like it should have accounted for different cultural attitudes towards stuff like honor, collectivism, egalitarianism, and hierarchy which might have affected willingness to accept unequal divisions.
I think the dollar amount was in the range of a few days or a couple weeks wages, not "enough to retire on" and definitely not "generational wealth". I agree with your expectations that for the much larger amounts, it should converge on accepting so long as the absolute amount offered is enough to really turn your head regardless of the total amount in the pool.
I would guess that the larger the pot of money on offer, the less likely that the game will be repeated. The sensible strategy (both game theoretic and instinctual) is very different for iterated game theory vs one-off games.
> Has anyone ever run a version of the Ultimatum Game with an actually important amount of *real* money on offer?
Yes, and it supports what you’re describing: https://web.stanford.edu/~alroth/papers/lihsug.pdf
From the abstract: ”rejections were less frequent the higher the stakes, and proposals in the high stakes conditions declined slowly as subjects gained experience.”
It’s interesting. If you look at figure 1c, you start seeing an increase in 33% offers as the stakes grow and players learn - and on the responder side, everyone stops rejecting these offers.
Why Muslims? Why not small business owners, or conservative Christians?
This sort of strategy only works if you're a group that has a reputation as angrily and self-destructively defecting if you don't get your way. The kind of people who will start wars they can't possibly win with their more powerful and better armed neighbours. The kind of people who will blow themselves up in crowded marketplaces for the joy of killing a handful of enemy children. The kind of people who'll fly passenger jets into skyscrapers just to gain a reputation as a group that shouldn't be fucked with, knowing that it will lead to a decades-long war that kills millions.
This kind of self-destructive defection is very much on-brand for Muslims and very much off-brand for everyone else. I would say that it often works out well for them, but that overall the costs they repeatedly incur from the many defections probably outweigh the benefits, which is why half the Middle East is rubble.
Isn’t it possible that they have their cake and eat it too by actually voting Harris but saying they won’t and didn’t.
It was confounding to see that all sides on the Israel/Hamas/Gaza war punished Harris and the Democrats and seemed to drift towards Trump. First blush would be to think the (D) positions and messaging were uniquely bad and contradictory and self-sabotaging. But the Harris campaign said nothing about anything, and Trump also said very little recently re: the Middle East.
I conclude that everybody is unhappy (to put it almost grotesquely mildly) with the war, and the party at the helm would be identified, even subconsciously, with the misery. It was probably the Right Move by all parties to say nothing much. Because what could one promise that people would believe? Whom could you champion that would not "Other" the rest of the electorate?
As Scott has noted, Trump has been a supporter of Israel and the Israeli government through all of his public life. Maybe there comes a time in conflicts where *certainty* is more valuable than specific outcomes. Maybe any sort of concrete action is better for all sides than chaos. Certainly Joe Biden, still US President, was not going to make any decisions. It is a strange result that Donald Trump of all people presents more gravitas to the world that the current US policymakers.
A quote for which I cannot remember the attribution, but can never forget the implication:
"...can't keep his feet out of the sh*t on both sides because he can't get the fencepost out of his a$$!"
Well I think there’s a strong case to be made that if Trump was president, Hamas would not have done 10/7 and so there wouldn’t be thousands of dead Gazans. Hamas strategy is based on the US not giving Israel a blank check to crush them since they can’t beat Israel militarily.
What is the strong case to be made? I don't see it. Its like the notion that 9/11 wouldn't have happened if Al Gore was President; why?
Again Hamas would have been too scared, since they know that Trump would give Israel a blank check.
Maybe. I doubt that, personally. what leads you to your opinion?
The way Trump behaved in his first term.
I don't buy the argument about people not choosing the "rational" solution in the ultimatum game.
What if the pot is $1b instead of a few dollars. Do you really think people will say no to, say $1m because of "fairness"?
I think this line of thought about what Michigan Muslims should do is completely wrong-headed. Most of the US is now so angry that people have no intellect left over to think about the pros and cons of voting strategy. The are desperate to scratch the anger itch, to get rid of the feeling that fuckwads who are dumb and wrong are doing things they hate and they have no power to stop them. Of course the present election is especially toxic and contentious -- though that may not change going forward, because a lot of the monstrous rage is created then aged in ideal conditions on social media, like monstrous wheel of limburger, . But in all situations where there are conflicting interests it's highly likely that some will be so overtaken by anger delusions that they are incapable of thinking about tactics, likelihood of various things, long term consequences of this and that, etc.
And the present state of ACX commenters is excellent evidence that rationalism does not protect people from having their mind overtaken by primitive anger, in which obviously ridiculous ideas seem true: Everyone who voted against my candidate is either evil slime or a moron. it is fine to mock them, insult them, etc. All of my opinions about my candidate are well-founded. I am able to make an accurate forecast of how things will go if candidate A is elected, and if candidate B is.
Not everyone who posts sounds like they have a case of that, but many do. Scott just gave somebody posting on the prediction market threat a brief ban, commenting that it's brief because everyone is crazy right now.
There's a book I'd like to recommend. It's the best corrective I've found for my own tendency to slide into Rage Idiot mode. It's *How to Think,* by philosophy prof Alan Jacobs. A better title would be *How to be Fair-minded*. This guy really understands the routes by which people end up believing and acting on toxic nonsense. He sees the process as driven mostly by efforts to manage self-esteem in situations involving ambition, desire to affiliate with an appealing group, fear of ending up in the lame group. .So he focuses on the emotions that are stirred in these situations, not on what constitutes clear thinking. I highly recommend it as a way of improving one's judgment and self-management in conflict situations.
> monstrous wheel of limburger
Nice ... image? Not quite the right word for an imagined smell... :-)
First, framing this as Michigan's Muslims "defecting" from the Democratic coalition, is predjudicial. Part of what the Democratic coalition has traditionally offered, and part of what made Michigan's Muslims join that coalition, was a commitment to human rights for oppressed peoples around the world. At very least, to e.g. complain to the United Nations when someone tries to perpetrate a genocide anywhere. And probably most Democrats have that something like thirty-five levels down from the top of their list of priorities, but presumably the Detroit Muslims care about the human rights of the Gaza Palestinians rather more than that.
From their point of view, it's the Biden Administration that has defected from the Democratic coalition, by basically signing off on fifty thousand dead Gaza Palestinians. And maybe that's the right thing for the administration and the party to do, for moral or geopolitical or domestic-political reasons. Sometimes coalitions have to cut loose some of their members to better serve the interests of the rest. But that doesn't mean it's wrong for the Detroit Muslims to believe that they are the wronged party here.
And as for what to do when you perceive that the other party has defected, I think that's pretty much straight out of Game Theory 101. If it's a single contest, you take the L and salvage whatever you can out of the matter, take the $0.01 in the Ultimatum Game or whatever. If it's an iterated game, you punish the defection and then renegotiate for the next round.
If Donald Trump is going to egg Benjamin Netanyahu on to literally exterminating the Palestinians, then this round is all there is, and you vote for the politician who just *might* try to stop that and will at least probably open the doors to more refugees than would Trump. If the present war is going to kill say 4% of the Gaza Palestinians and then we'll have to deal with the reconstruction after this war, and then the next war and the reconstruction after that, then maybe it's worth enduring four years of Trump for a chance at having a reforged Democratic coalition that takes their concerns seriously going forward.
And it's tempting to extrapolate from there to "see, this proves that all those whiny activists don't *really* believe all that genocide nonsense!". But really, all those whiny activists are neurotypical human beings rather than nerdy rationalists, so they don't do the math for Game Theory 101. They just implement long-established heuristics, which are basically "everything is iterative, punish all defectors, we've hardwired your brain so it feels *good* when you punish defectors, and surely you can come up with the case-specific rationalization".
> the longest Discord thread in the history of Project Lawful
Good times. :-)
And yes, it's important to understand your goals. Are you trying to achieve an equitable balance of power in a coalition, regardless of how little actual power that might give you? Or are you trying to maximize your short term power in the coalition at the expense of long-term interests like avoiding backlash? (One must note that, in the case of Palestinians, there's a war going on.)
I think a good choice is to demonstrate power and control by cementing a voting block that acts in a coordinated way. Announce the day before that your people will vote for "Caliph Haroun El Poussah", or Cornel West, or someone else neutral who won't win, and then let the vote totals show the outcome. Eventually someone will want to woo your block, or not. Another complication is that that's negotiating in good faith, which is a quokka-like trait that may not have any place in politics, even if an ASI might appreciate it.
As for how much to ask for, that's where the Shapley value comes in, but like you say, it's important to remember that the coalition involves a lot of other agents. If your goal conflicts with the goals of more important coalition members, then you're kinda screwed. On the other hand, it's sometimes possible to negotiate directly with other coalition members, and get them to change some of their goals that are less important to them, to match yours, while you adopt some of the goals important to them. (As suggested in Eliezer's "Three Worlds Collide".)
Harking back to the "threat" discussion, it's dangerous to take actions that harm your own interests, as a minor player in a multi-agent system. If there's an R coalition and a D coalition, and you want to join the Ds, and the Rs aren't ever going to help you, there's always a chance that harming your interests and the D's interests will result in an R victory. The D coalition, being bigger and stronger, is probably better placed to recover than you are. (Because it's not just a competition for abstract points; political losses can actually damage your ability to act in the future.) Plus it pisses people off, and they may decide to let you keep hurting yourselves.
Personally, I usually take a low-brainpower way out. In races where I don't actively support one candidate, and where it's predicted to be a blowout, I vote for the underdog. That makes the race more competitive. Now that I've articulated this in public, though, I should point out that Kant's Categorical Imperative has some things to say about coordination, and so I wouldn't recommend that the entire world blindly adopt this low-brainpower policy.
This all assumes as starting point that US Arabs
- are primarily motivated by Israel
- and that motivation is that they want to punish it.
Both of these strike me as not obviously true. They are what the Dems want to be true because that fits into the rest if their theory of the world, and doesn’t require them to rethink anything, but that doesn’t actually make them true…
I’d put the Arabs more in the same bucket as Hispanics — utterly sick of being defined by their ethnicity above all else, and to add insult to injury, having that ethnicity defined by people who thinking reading Said makes them experts on how you think.
"Is there some level of Democratic corruption at which I should vote Republican to “punish” the incumbent even if I think the Republican would be a worse leader overall? I’m not sure, and currently lean towards no, but I can’t say it doesn’t tempt me."
There should be, assuming the Republican has good character (i.e. a John McCain type Republican, who has good character but horrible (to me) policy). Because a person of good character will research, experiment, and reassess to ensure best outcomes are met, not be blindly ideological. For example, this is why Pete Buttigieg was such a great Secretary of Transportation: he kept looking for the best ideas to achieve his goals, and found them, even when they were against his initial positions.
When I see someone say: "Let's all vote in X way, and I acknowledge that this looks like it's contrary to our interests but here's my argument for why we should do it anyway," my first thought is to check whether the speaker is actually a member of my coalition (as opposed to a member of the other side engaged in creative vote-gathering).
The biographies at abandonharris.com look pretty real to me, so in this case I'm not seriously suggesting this was a trick. But I do have to wonder how long we have before it becomes possible to generate a "Movement of X Interest Group Voting For Y Candidate" page algorithmically.
I think voting third party for someone you really would like to be president more is the answer here; it maximally hurts Kamala, but since you're not voting to hurt yourself, I don't think it counts as a threat.
> I still wish I had a better idea of when to protest-vote [...]
Vote with your feet. (Or threaten to do so.)
Relevant: https://x.com/RichardHanania/status/1762275726215258354 ... I guess there though, you can't do anything to pacify the scorpion.
Anyway, perhaps these Muslim voters don't want dead Gazans more than anything else. I think Hamas would have been too scared to do October 7 with a Republican president. So, no dead Gazans.
Why are Muslims against Israel though? Who else are they against?
I think trying to apply analysis from perfect-information games with rational players to politics is generally a mistake, and that working out a correct strategy probably involves a deeper understanding of the psychology not just of the people you're ultimately trying to influence, but of the people in your own movement that you're trying to persuade to adopt your preferred tactic, than most people - definitely including me! - have.
"or even punish them by veering further anti-Israel?" do you mean pro-Israel here?
The game theoretic approach only works if the other players also follow game theory. eg: beginners luck in poker.
Politics works on shock value. 'The muslims voted for Trump' drives better shock value than any of the alternatives. Therefore, you vote for Trump. Plus, 1 vote is a 2 vote swing if it goes Kamala -> Trump, but a 1 vote swing from Kamala -> 3rd candidate. Always maximize your power. Go Kamala -> Trump.
However, shock value only serves to draw attention. And attention isn't always good. Attention helps highlight a show of power, only IF you have power.
The way the math played out this election, PA was always going to decide the outcome. If Trump won PA, none of the other swing states added up for Dems. If Kamala won PA, she could lose 1 swing state and would've likely swept the rest. Plus, Trump & Obama have showed that people vote based on a national vibe. Since the 90s, PA-Wisc-Mich have all swung in the same direction. The mid-west [1] / great-lakes mega-region [2] swings together.
From the attention drawn by Dearborn, Dems are going to realize that American Muslims don't matter. This time, Dems sat on the fence by allowing campus protests to get really bad and rejecting Shapiro as the obvious swing-state VP candidate. Clearly, it wasn't enough for Muslims.
Non-black (NB = not descendants of slaves) Muslims are 1 % of the US population, and Michigan is the only swing state with a substantial NB Muslim population (2%). So at best, a perfect pro-democrat Muslim vote base can help make a 2% difference in Michigan. That's pretty small. Especially when it is traded off against the most important state (PA), which has 3.5% Jews in comparison.
Going off attention and power as the 2 metrics of interest. Here's a better decision tree for protest voting an already lost election :
1. If in_state_power can flip future national election -> protest vote such that it draws most attention = vote Trump.
2. If in_state_power cannot flip national election -> vote to create narrative around power and minimize attention = vote Dems.
In scenario 2, all michigan muslims vote Democrat. At a 1.4% difference, it is a swing-able number for NB michigan muslims (2%). Kamala still loses nationally, but she wins 1 and only 1 swing state -> Michigan. New narrative : The country abandons Kamala, but the Muslims pull out unexpected win.
Now the same 1.4% creates a greater impression of power in this new narrative. The narrative doesn't get much scrutiny because it falls in line with existing preconceptions (muslims matter, muslims vote dem). In 2024, the Dems gave NB-muslims more leeway than their demographics demand. NB-Muslims might think that this was not enough for them, but their inflated entitlement doesn't match their demographic power. The best thing to do, is to let the status quo continue and lay low. I predict, they will regret drawing attention to themselves.
I'm getting ahead of myself, but 2028 will be about the collective will of the 2 main mega-demographics: the great-lakes mega-region & Latinos. If you aren't part of either mega-demographic, you don't matter.
[1] https://fivethirtyeight.com/wp-content/uploads/2014/04/hickey-map-midwest2.png
[2] https://upload.wikimedia.org/wikipedia/commons/8/8d/Emerging_US_megaregions_with_cities_labeled.png
I did a lot of ruminating on this back in the day, when I was a little bit more active trying to help get gay rights legislation passed in the US. The DNC was *awful* about actually putting their own resources behind advancing gay rights, to the point that Clinton in the 90s was actually signing anti-gay laws like Don't Ask, Don't Tell. But obviously we still had to keep voting for them, because even though they were merely the ones punching us slightly less, the alternative was "actually homosexuality should be illegal". Even when it started being a wedge issue on the other side, such that running on gay rights got them more votes than it cost, they still basically let the courts do most of the heavy lifting. To this day, very few laws that you could call "pro-gay" have ever gone through the federal congress, most of them basically symbolic.
But unfortunately, I don't think there's really a viable alternative to "vote for the guys passing slightly fewer laws banning your community from public life, then try to route around formal politics and hope you can change the culture." I mean, it's left me feeling permanently rather burned by politics, but I don't think the DNC has to actually care about that? Without some viable BATNA, they can suck as much as they want. As long as there's an effective duopoly on governance, these parties pretty much have us by the short hairs.