Against opposing SJ activism/cancellations

I disagree. It seems to me that the EA community's strength, goodness, and power lie almost entirely in our ability to reason well (so as to be actually be "effective", rather than merely tribal/random). It lies in our ability to trust in the integrity of one anothers' speech and reasoning, and to talk together to figure out what's true.

Finding the real leverage points in the world is probably worth orders of magnitude in our impact. Our ability to think honestly and speak accurately and openly with each other seems to me to be a key part of how we access those "orders of magnitude of impact."

In contrast, our ability to have more money/followers/etc. (via not ending up on the wrong side of a cultural revolution, etc.) seems to me to be worth... something, in expectation, but not as much as our ability to think and speak together is worth.

(There's a lot to work out here, in terms of trying to either do the estimates in EV terms, or trying to work out the decision theory / virtue ethics of the matter. I would love to try to discuss in detail, back and forth, and see if we can work this out. I do not think this should be super obvious in either direction from the get go, although at this point my opinion is pretty strongly in the direction I am naming. Please do discuss if you're up for it.)

Someone wrote this to me privately. I agree with the substance of the criticism and has since edited the post accordingly.

> As a general strategy, it seems much better for most people in the community to [...] quickly disavow any associations that could be seen as potentially problematic.

This part seems objectionable to me even if I agreed with the rest of your post.

1. Public disavowal can increase the chance that the accused person will suffer unjust bad outcomes. This starts to slide away from ‘don’t protect your peers from being burned as witches’ to ‘preemptively help burn your peers as witches so you don’t get burned yourself’. This seems like bad game theory to me, so I question its wisdom from a selfish perspective. And it seems especially bad altruistically, if the people you’re helping burn are fellow EAs.

2. If you don’t think the accused person is actually a witch, then helping burn or condemn them also seems like a violation of some of the important ethical principles that helps people coordinate.

If I expect my peers to lie or stab me in the back as soon as this seems useful to them, then I’ll be a lot less willing and able to work with them. This can lead to a bad feedback loop, where EAs distrust each other more and more as they become more willing to betray each other.

Highly knowledgeable and principled people will tend to be more attracted to groups that show honesty, courage, and integrity. There are a lot of contracts and cooperative arrangements that are possible between people who have different goals, but some level of trust. Losing that baseline level of trust can be extremely costly and cause mutually beneficial trades to be replaced by exploitative or mutually destructive dynamics.

Camaraderie gets things done. If you can create a group where people expect to have each other’s back, and expect to be defended if someone lies about them, then I think that makes the group much more attractive to belong to, and helps with important things like internal cooperation.

But even absent camaraderie, basic norms of civil discourse get an awful lot done too. Norms like ‘we won’t help make things worse for you and spread misinformation about you if someone is unethically targeting you’ get you a lot, even if you lose the valuable norm ‘we’ll defend you if someone is unethically targeting you’.

3. Another problem with denouncing people who you don’t think deserve denunciation is that it puts you on the record about any person, group, or idea anyone ever wants to make you publicly weigh in on. If you refused to participate in the witch-hunting as a matter of principle, then this might lose you some reputational capital in the near term, but in the long term it would make it harder for people to infer ‘oh, so you do endorse this other thing’ from your decision not to disavow something later.

One way of thinking about the free speech meme, 'though I disagree with what someone says, I'll defend to the death their right to say it', is that it’s functioning as exactly this kind of game-theoretic strategy right now. On this way of doing things, people get to avoid condemning each other as witches in academia - in fact, they even get to actively work to help and protect each other, in cases where they think the accusations are unjust, harmful, or false - all without ever endorsing or disavowing any of the actual positions under discussion.

To the extent this works, it works because a large group of people has agreed to an explicit strategy of protecting people even when they disagree with or dislike them. This lets you protect the falsely accused (or at least avoid accusing them of witchcraft yourself) without going on public record about every accusation that’s currently blowing up on social media.

This strategy doesn’t make them immune to cancelation, but especially so long as the strategy is widespread, it provides massively more leeway and protection against discourse evolving toward Red Scare dynamics.

We needn't take on reputational risk unnecessarily, but if it is possible for EAs to coordinate to stop a Cultural Revolution, that would seem to be a Cause X candidate. Toby Ord describes a great-power war as an existential risk factor, as it would hurt our odds on: AI, nuclear war, and climate change, all at once. I think losing free expression would also qualify as an existential risk factor.

First they came for the Communists
And I did not speak out
Because I was not a Communist

Then they came for the Socialists
And I did not speak out
Because I was not a Socialist

Then they came for the trade unionists
And I did not speak out
Because I was not a trade unionist

Then they came for the Jews
And I did not speak out
Because I was not a Jew

Then they came for me
And there was no one left
To speak out for me

I think it's important to consider the general principles in question even if the particular instrumental claim 'defending accused witches doesn't do as much good, as you would in expectation be prevented from doing via your work on slavery if you defended accused witches.'

This seems to imply some general principles which don't seem that attractive, i.e. "Don't speak out against/defend against/protest one injustice if you think it will get in the way of working on injustices you care about more.'

This seems like the kind of violation of commonsense morality in the name of utilitarian instrumental goals that the EA community generally warns against. (I also worry that this specific violation of normal moral obligations like 'defend the innocent' 'speak the truth', makes it more likely that people will generally violate such norms in pursuit of their utilitarian goals).

This stance also seems quite shaky, since it seems like we would not generally support such reasoning if the cases were changed just a little bit e.g.:

"We should not speak out against slavery, because it would get in the way of our important anti-poverty work."

"We should not defend or associate with controversial _racial justice activists_, because it will reduce our other EA work."

This also seems bad from a reciprocity standpoint i.e. if slavery activists don't defend or associate with witch defenders, then witch defenders, by the same token may not defend or associate with slavery activists (and so on for other controversial groups). These reciprocity considerations might apply either directly and instrumentally or indirectly via defending the general norm.

Your position also seems even more extreme than how I described it above at points, i.e. "it seems much better for most people in the community to watch what they say in public somewhat, _be careful with their public associations_, and _minimize public contact with any associations that could be seen as potentially problematic_." This goes beyond merely not publicly defending groups. Add "minimiz[ing] public contact" with the groups I gave as examples above and this position seems even more problematic.

That said I think one part of your somewhat concessive, but somewhat ambiguous final paragraph is potentially true:

Individuals can do so... but doing so as a group is a dangerous correlated risk to the movement.

I think it's good to grant that individuals can stand up for accused individuals. I still think that a statement warning off EAs "as a group" is potentially problematic, because this could mean "It's OK for a small number of EAs to do this but not too many", which seems as objectionable as "It's OK for a small number of EAs to publicly oppose slavery, but not too many." But if "as a group" meant "The EA community shouldn't make official public statements as a whole on the political debates of the day or on other controversial issues, and nor should official EA orgs' (which I don't think was your intended meaning), then I would agree with this principle.

Someone said this to me recently:

I think the pace of cancellations and attempted cancellations has been picking up steam and has gotten pretty damn fast very recently. It is possible I am not sufficiently familiar with base rates, but it does seem like the current pace is quite high, so either it has accelerated or people on the peripheries of our community are unusually likely to be targeted. In either case this is cause for concern.

Eg, REDACTED1's firing, REDACTED2 asked to resign his VP position (but he still gets to be faculty), petition to remove REDACTED3 from his honorary titles.

Also lighter stuff like REDACTED4 forced to delete his Twitter account, and different spats within newsrooms, that I'm not following as closely.

In either case, I've updated towards being much more cautious than usual. There are already things I don't talk about, but I've updated to putting "nuanced takes on social justice topics" as on par with "criticizing the Chinese government as a Chinese citizen" for things I don't publicly talk about.

I talked to other people who are online a lot, and some current college students. The general consensus was substantial concern/alarm, and that people who are older and/or less aggressively aligned may have missed.

Coordination infrastructure that would allow sane people to defuse runaway information cascades in generality would be very valuable to discover, unless physics doesn't allow it.

Can any historically and/or sociologically familiar people comment as to the ability for the witch hunts to have been stopped by counterfactually motivated and capable parties, and what order of magnitude of motivation and capability might have stopped it?

I received this as a private message. It was unclear to me if it was initially intended as a comment, but I asked and they gave permission for me to do this:

I'm quite bothered by the implicit assumption that this is in fact a cultural revolution. I think the degree to which people will find supporting Hsu offensive is completely dwarfed by how offensive people will find offshoots of the Floyd protests as equivalent to the Cultural Revolution. As noted in https://www.scottaaronson.com/blog/?p=4859, there are steel man versions of the Hsu complaint in particular (though I don't personally have a view).

So if you're trying to preserve EA credibility, implicitly saying we're living through a Cultural Revolution could be extremely damaging.

Should be "is completely dwarfed by offensive people will find the idea that offshoots of the Floyd protests are equivalent to the Cultural Revolution".

One can defend or not defend Hsu (again, I haven't looked at it enough to have a personal view) without making extreme claims about Cultural Revolutions, and both routes (defending or not defending) seem *way* better than crying Cultural Revolution.

If it isn't clear, I'm biased here, in that I personally find the Cultural Revolution claims offensive.

I may have just missed this in the comments below, but FWIW: On top of all the other points that have been made in opposition to this stance, I would also assign very low credence to the implied claim "if we don't [do things that oppose cancel culture], then we'll be able to avoid getting canceled during the 'cultural revolution'." I would suspect that if this "cultural revolution" (which I already consider implausible) were nearly as bad as you suggest, EA as a movement would already get targeted regardless (especially if it's the case that the whole movement will be held collectively guilty for a subset of the movement speaking out about something), and thus it would have an even smaller fractional expected value. To clarify further, here the witch analogy that you use is potentially misleading because with witch hunts the scope is at least ostensibly limited to the instances of "witches." This could of course be expanded to include "witch sympathizers", but it's at least more plausible that by avoiding getting involved one can continue their abolition work. If however the witch hunt were to grow into an entire philosophy that says "anyone not primarily concerned with finding witches and burning them will be treated as witch sympathizers," then you face the lose-lose situation (darned if you do, darned if you don't).