Comments
I'm skeptical of metrics like "x% of people involved said they were satisfied" for estimating cost-effectiveness. Customer satisfaction doesn't really connect very well to any of the things I care about; in most cases I'm happier with a rough estimate of lives saved/units of suffering prevented/QALYs purchased/etc. per dollar than with a more precise accounting of things that touch less directly on the end goal.
I was extremely disappointed to see this tweet from Liron Shapira revealing that the Centre for AI Safety fired a recent hire, John Sherman, for stating that members of the public would attempt to destroy AI labs if they understood the magnitude of AI risk. Capitulating to this sort of pressure campaign is not the right path for EA, which should have a focus on seeking the truth rather than playing along with social-status games, and is not even the right path for PR (it makes you look like you think the campaigners have valid points, which in this case is not true). This makes me think less of CAIS' decision-makers.
Ya, really sad to hear that!
I mean, if they were going to fire him for that, maybe just don't hire him. Feels kind of mercurial that they were shamed into caning him so easily/soon.
Ofc, I can understand nonviolence being, like, an important institutional messaging northstar lol. But the vibes are kind of off when you are going to fire a recently hired podcaster bc of a cringe compilation of their show. Seriously, what the hell?
I do not think listening to that podcast made me more violent. In fact, the thought experiments and ideation like what he was touching on is, like, perfectly reasonable given the stakes he is laying out and the urgency he is advocating. Like, it's not crazy ground to cover; he talks for hours. Whatever lol, at least I think it was understandable in context.
Part of it feels like the equivalent of having to lobby against "civilian nuclear warheads". And you say "I wish that only a small nuclear detonation would occur when an accidental detonation does happen, but the fact is there's likely to be a massive nuclear chain reaction." and then getting absolutely BLASTED after someone clips you saying that you actually WISH a "small nuclear detonation" would occur. What a sadist you must be!
This feels like such stupid politics. I really want John Sherman to land on his feet. I was quite excited to see him take the role. Maybe that org was just not the right fit though...
Hard to know the full story but for me this is a weak update against CAIS's judgment.
Right now CAIS is one out of a total of maybe two orgs (along with FLI) pushing for AI legislation that both (1) openly care about x-risk and (2) are sufficiently Respectable TM* to get funding from big donors. This move could be an attempt to maintain CAIS's image as Respectable TM. My guess is it's the wrong move but I have a lot of uncertainty. I think firing people due to public pressure is generally a bad idea although I'm not confident that that's what actually happened.
*I hope my capitalization makes this clear but to be explicit, I don't think Respectable TM is the same thing as "actually respectable". For example, MIRI is actually respectable, but isn't Respectable TM.
Edit: I just re-read the CAIS tweet, from the wording it is clear that CAIS meant one of two things: "We are bowing to public pressure" or "We didn't realize John Sherman would say things like this, and we consider it a fireable offense". Neither one is a good look IMO.
I have no idea about this particular case, but there are difficult comms tradeoffs to be made.
If your are working for an advocacy org, sometimes it might be important to make messaging compromises to build alliance and get policy through. I can see a good argument that a public comment like this could make others work at the org a lot harder, so it might not be the best job for him in that case. People need to be saying stuff like he said, but maybe in a different org?
Truth seeking is important. So is building alliances.
I regret that this has happened, but there has apparently just been a confirmation of what I said four days ago about how incendiary such comments could be from the point of view of people who are emotionally troubled, and how there is a real risk of physical harm to innocent people. These are not "social status games", and statements to that effect are irresponsible.
The comments quoted in that tweet come perilously close to an incitement to violence. If you don’t think that anyone would actually commit violence due (partly) to ideas related to the rationalist community or AI alignment, I’ll point out that this has already happened, and possibly up to six people are dead because of it. (That’s not the whole explanation, but it’s part of it.)
When you are speaking publicly, I think you have a responsibility to be extra cautious not to incite violence or say things that could be interpreted by someone as encouraging violence. You’re not just talking to the median person listening, you’re talking to everyone listening, including people who are impressionable, emotionally unwell, and may be predisposed to violence.
It’s unfortunate, but this is what it means to hold a position of responsibility in our society. You have to consider these things. I really strongly disagree with dismissing these as mere “social status games” — we are talking about a real risk of irreparable physical harm to innocent people here, even if small.
What’s strange here is the “warning shot” already happened — up to six murders, perpetuated by people who almost certainly had multiple risk factors going on (as almost always seems to be the case), but for whom the discourse, ideas, and subcultural social norms of the whole LessWrong/Bay Area rationalist/AI alignment world seemed to play a role.
[Edited on Nov. 22, 2025 at 6:00 PM Eastern to add: disturbingly, what I said in this comment turned out, just a few days later, to be more prescient than I could have realized.]
They don't come remotely close to it. That's a wildly dishonest characterization.
No, it’s an incredibly accurate characterization.
I don't think you're familiar with what actually constitutes incitement to violence.
You have to look beyond what was literally, directly said to what the most extreme people who are listening to those remarks might infer, or might feel encouraged to do. Saying that people should burn down AI labs and the employees should be jailed for attempted murder is not literally, directly saying someone should commit violence against the employees of AI companies, but it is easy for someone who is in an extremist mindset and who is emotionally unwell to take what was said that extra step further.
And there is no point arguing about what is theoretically, in principle true or not when this violence is already happening, or being threatened.
No, actually; the legal standard used in courts is what was actually said, which needs to include a clear call for violence to be carried out at some point in the near future. It's extremely frustrating to me that you're misusing legal terms to lend your arguments weight they don't hold; please cease to do so. https://en.wikipedia.org/wiki/Brandenburg_v._Ohio contains plenty of helpful information if you'd like to learn more about what "incitement to violence" means in America.
I didn't say it was an incitement to violence, I said it was perilously close to one. What that means is that the person making such statements can, indeed, completely avoid legal liability for such statements, and can plausibly deny any moral responsibility if any violence occurs, although the actual effect on a very small minority of people listening — who aren't in a headspace where they can safely process these kinds of inflammatory proclamations — might be, plausibly, to encourage violence.
The important question is not what kind of speech is illegal or not, the important question is what kind of speech might be taken as encouragement (or discouragement) of the kind of violence or threatened violence that just happened, whether or not that is the speaker's intention. I'm not trying to make a claim about what's illegal or not, I'm making a claim about what kind of public statements are responsible or irresponsible.
It's not "perilously close", because it's very different from incitement to violence. I have explained that incitement to violence requires a call for violence which is time-scoped to the near future; Sherman's statement did not include a call for violence at all. You are correct that he bears no moral responsibility for the actions of people who heard his statements.
"Perilously close" has no legal definition, so what you are asserting is a matter of opinion, not a matter of fact. My intention in using "perilously close" was to convey that such statements have a similar kind of danger to statements that would meet the legal definition of incitement to violence, even though they are perfectly legal.
You know that I did not say people who make such statements bear no moral responsibility for how their words are interpreted, so I'm not sure what your intention is in making that false statement.
Since you have not signaled good faith, I won't engage further.
They do not have a similar kind of danger; you are making false equivalences. Thank you for ceasing your censorious fearmongering.