This can’t be right, because Bob is supposedly rational and one of his two choices is irrational.
This seems like it says the opposite of what you are saying? It's saying directly that Bob will choose defect if he is rational. Indeed, he says that the very premise of the game is made impossible by Bob's rationality, because defection is the only rational choice.
I think you are just summarizing this section incorrectly. Claude also says:
No. Binmore argues the opposite. In this section he calls the twins argument a fallacy, and he says Alice should not reason her way to dove on the grounds that Bob will mirror her choice.
His main points:
- The argument is a fallacy. It assumes that because both players are rational and face the same problem, they will necessarily choose the same action, so Alice should pick dove, since (dove, dove) beats (hawk, hawk) for her.
- It answers the wrong game. The reasoning would be right for genetically identical twins whose behavior is fixed by their genes. But that is a one-player game, not the Prisoner’s Dilemma.
- The Prisoner’s Dilemma assumes independent choice. Bob’s choice can’t be assumed to match whatever Alice picks. Bob is supposedly rational, and hawk (defection) is the rational choice, so dove would be irrational for him.
- Weakening the assumption doesn’t help. If you only assume the choices coincide with high probability, you’ve still made them correlated, which means they aren’t independent, so it isn’t the Prisoner’s Dilemma anymore.
- Correlated equilibrium doesn’t rescue dove. Even with Aumann-style correlated information, hawk is strongly dominated, so rational players still wouldn’t play it...
One correction to that last point. Binmore’s wording is that they “still wouldn’t play hawk” because hawk is strongly dominated. In his labeling here, hawk is the dominated strategy, which means the passage is using hawk for the cooperative-sounding action in its framing of dominance (this is a quirk of the excerpt’s hawk/dove labeling). The substantive upshot is the same: he holds that rational players play the dominant strategy in the one-shot Prisoner’s Dilemma, and he rejects the twins argument as a route to cooperation.
I should flag that the last sentence is inconsistent with the earlier ones, in which dove was the cooperative move and hawk the “rational” one. As written, the excerpt seems to have hawk/dove swapped in that final line, so the passage looks internally muddled there, and it’s worth checking the original text. His overall position on the twins argument is clear regardless.
And indeed, this is the only way I can make sense of this passage. I think the last sentence genuinely is just flipped and the author made a typo. "Dove is strongly dominated" is the only justified claim you could make in a classical game theory context. Hawk is definitely not strongly dominated, or at least it would require perfect correlation (in which case I agree we are not playing the twin DP at all anymore). I think you ended up misreading this section because of a typo.
Summary: Scott Alexander’s Against SURA post is arguing against a strawman version of unilateral pauses. He admits as much deep into the post. I think his conclusions fail against stronger pause plans for conditional or limited unilateral pauses, where training resumes if the pause does not cause others to pause, and where personnel are redirected toward work on alignment and product, rather than simply fired. In addition, Scott’s post is incorrect about the simplicity of the game theory involved: getting companies to agree to conditional pauses is not strictly easier than convincing them to pause, as the former requires companies to (promise to) take actions based on ambiguous, game-theoretically-informed evidence. That is not the kind of evidence companies tend to act on.
The good version of unilateral pausing is good
Here is the good version of unilateral pausing:
There are multiple places in the post where Scott makes it clear he wants the reader to be thinking of much worse plans than this. Emphasis added:
Again, the correct version of a unilateral pause is a company pivoting to safety + security + product, not “disbanding their entire company immediately”.
This is again arguing against a strawman: OpenAI would not be “out of the industry” if they decided to unilaterally pause. Sam Altman loves to build products! He would very much still be in the industry.
It is possible that Scott will correctly point out that some people do advocate for the extreme version here, but I don’t think he should get credit for this. The title of the post is “Against Stop Unilaterally and Raise Awareness”. I think it is perfectly reasonable for someone to say that SURA-the-phrase applies to my better version. His post is very strongly being interpreted as arguing against unilateral stopping in general. It is being interpreted that way because of his title and framing.
Scott thinks the good version of unilateral pausing is good, or at least exempts it from criticism
Scott appears to agree that the good version of unilateral pausing is good, but attempts to claim that only one group is advocating for this good version, the AI Futures Project.
This should not be a parenthetical. If the goal is clarification for people who haven’t thought about the details much, and Scott believes as I do that there is a good version of unilateral pausing, it should have been in the opening section. It is not. The reader should not be expected to review every section before realizing that the title of the post is wrong.
Oversimplified game theory is bad, and you should feel bad
The core argument behind unilateral pausing is that the game theory is complicated, and it is better to go via deontology and other non-consequentialist analyses if you can’t model the game theory well. Scott reasonably points out that one shouldn’t just say this: he wants details. Unfortunately, in the absence of others providing details, he provides his own game theory and makes basic mistakes, providing evidence for the original why-deontology point.
In particular (emphasis added):
Agreeing to pause conditional on others pausing is not strictly easier than unilaterally stopping. There are three reasons for this: (1) “some definition” hides dragons, (2) promises for future actions can be harder than actions, even if the immediate actions are stronger, (3) the boundary between frontier lab and not is ambiguous.
“Some definition” hides dragons
There are several versions Scott could have meant by “some definition”:
It should be clear that getting a company to agree to the general version is out of the question: it is too wide a promise. Set that aside.
The internal definition version is useless without additional rounds of negotiation. If all companies come up with their own definitions and agree to act on them if others act on their own internally-generated definitions, the immediate outcome is no action. Additional rounds of negotiation then clearly falsify “strictly easier”.
The external definition version is also not strictly easier. It is in practice much harder for a company to agree on an external definition than to take action based on their own internal definition. This is a combination of not-invented-here syndrome and different definitions affecting different companies in different ways due to different team structures and research programs.
I am not guessing here. I have spoken directly to multiple people at all of Anthropic, OpenAI, and DeepMind about unilateral pauses, and they are all sketched out by this question of the precise definition of a pause. One of those people was the CEO of one of those labs. Hiding this complexity behind “some definition” is oversimplifying the game theory, and that results in bad analysis and bad conclusions.
Now of course, unilateral pausing also means definition mismatch and multiple rounds of negotiation. It might be the case that one company pauses according to definition D1, and another company refuses to follow suit unless they strengthen to definition D2. But accepting that we live in a multiround negotiation world is one of the core arguments behind unilateral pauses: the goal is to start a process that can then ratchet up through negotiation.
Promises can be harder than actions
It is not the case that promising to do X in the future based on condition Y is strictly easier than doing X now. This is because
The above are some abstract first principles. But they conform to how AI company executives speak about the issue, and how I’ve seen negotiations break down in the past: Executives are very uncomfortable making conditional promises.
As one example, DeepMind’s attempt to become a company limited by guarantee (CLG) failed in part for similar “reality fails to conform to simplistic game theory” reasons. DeepMind leadership was attempting to guard against “AI capabilities are crazy” worlds where risks are high, and the Google Board was attempting to guard against “AI fizzles and DeepMind wastes a bunch of money” worlds. Simplistically, there was an arbitrage to be made. In practice, negotiations fell through. One should be very careful when assuming that simple theorems of simple game theory apply.
How many frontier labs are there?
Another reason it is not strictly easier to do a Scott-version conditional pause is that it is not clear how many frontier labs there are. In the U.S., the number is somewhere between 2 and 6. There are clearly worlds in which getting 2 companies to do a unilateral and then bilateral action is easier than to get 6 parties to do even a much weaker action, and I claim we are in one of those worlds.
Definitional fairness is less important than survival
As mentioned, when I talk to people at AI companies, they are very concerned that if they stop, other people will stop slightly less, and it will be unfair. We can contrast the following two risks from AI:
The second risk is more important!
Now, one can argue that the first risk is important too, and indeed significant work should go into minimizing the unfairness, since that unfairness does impact pause effectiveness. Too much unfairness means we don’t slow down AI capability progress enough and don’t buy enough time. But I think that fairness is to be achieved over time, after multiple rounds of action, and should not be a precondition for taking the first step.
It may be much easier for an AI company to act on their own definition than by one imposed on them externally. So shifting the conversation to the relative unimportance of the definition used for the first action may be critical to actually causing that action to happen. This applies not just to unilateral or bilateral pauses but also to the grand international coordination that Scott prefers. In that context, it is notable that Scott’s only engagement with definitions is using the phrase “some definition”, handwaving a bit, and adding a footnote about forecaster wording.
Once we will fill in the subtleties around definitions and the details of the game theory, the correct action is to stop.