“If you really believed X, you would do violence about it. Therefore, not X.”
I’ve seen this a few times with X := “AI extinction risk” (maybe fewer, seems like something I’d be prone to overestimate)
This argument is pretty infuriating because I do really believe X, but I’m obviously not doing violence about it. So it’s transparently false (to me) and the conversation is now about whether I’m being honest about my actual beliefs, which seems to undermine the purpose of having a conversation.
But it’s also kind of interesting, because it’s expressing a heuristic that seems valid—memes that incite “drastic” actions are dangerous, and activate a sort of epistemic immune system in the “uninfected.”
In this particular case, it’s an immune disorder. If you absorb the whole cluster of ideas, it actually doesn’t incite particularly drastic actions in most people, at least not to the point of unilateralist violence. But the epistemic immune system attacks it immediately because it resembles something more dangerous, and it is never absorbed.
(And yes, I know that the argument is in fact invalid and there exist some X that justify violence, but I don’t think that’s really the crux)
As an aside, I think a lot of “woke” ideas also undercut the basic assumptions required for a conversation to take place, which is maybe one reason the left at its ascendancy was more annoying to rationalists than the right. For instance, explaining anything you know that a woman may not know can be mansplaining, any statement from a white person can be overruled by a racial minority’s lived experience, etc., and the exchange of object-level information becomes very limited. In this case also, there is a real problem that wokeness is trying to solve, but the “solution” is too dangerous because it corrosively and steadily decays the hard-won machinery underlying productive conversation.
I think the underlying (simple) principle is that a pretty high level of basic trust is required to reach agreement. That level of trust does not exist on most of the internet (and I hate to see how much of the conversation on lesswrong now links to twitter/X, where it clearly does not exist).
As a “leftist” person, agreed! We’re all pretty fragmented. I’m working on creatively writing about these dynamics through the medium of a narrativized self. Sometimes the message doesn’t land as intended but all we can do is keep on trying to connect on the basis of shared consciousness (and faith in the goodness of everyone at their core).
“If you really believed X, you would do violence about it. Therefore, not X.”
I see this more often in areas other than AI risk, but regardless of context I agree it is a weak argument. It requires multiple other premises also be true. Let’s assume that people are not going to be moved by pacifism or violence is always wrong and just skip to practicals:
The violence one is capable of would cause a meaningful change in direction or time until onset of X
The targets of violence one has access to, are responsible for, or able to influence X
There will be no backlash to the violence that increases the likelihood of X
And these are just the basic ones that would rule out most reasons for violence.
If someone thought they could get away with eliminating a couple CEOs and data centers, would their belief in extinction risk justify doing it? Do they think it would change timelines? Because I don’t think it would. I give it below 1% that the loss of all the c-level execs at all frontier labs stops AI development. It jumbles up who gets there first, it changes who is personally taking what actions. But it doesn’t have a high chance of saving anyone, as far as I can tell. So bullet one applies. Also, I am quite confident that this would make the AI safety advocates lose ground, be portrayed badly in media (classic and social both), and thus it would more than 50% likely work against the goal of stopping AI risk (bullet 3), not for it.
There are very few people for whom violence is a realistic option. Because of these unstated premises.
The Gen Z riots globally , by some metrics have been doing pretty good. For example, the Wikipedia page claims 3 fails, 7 successes, and 14 ongoing with 4 of those alt least partly successful. This tallies to multiple governments overthrown, and several corrupt regimes having to back down on measures for self-enrichment. However, the new regimes appear to be just as corrupt, but favoring a different group. If anti-corruption was the goal, then bullet 1 applies and violence isn’t helping. If the goal is “the other team loses” then perhaps the Wikipedia claim holds up.
AI safety seems to be in a similar scenario where the most likely outcomes of violence are at best changing who gets their first, not whether we get there at all. I don’t see this as justification for violence, and to me the original claim is not valid.
“If you really believed X, you would do violence about it. Therefore, not X.”
I’ve seen this a few times with X := “AI extinction risk” (maybe fewer, seems like something I’d be prone to overestimate)
This argument is pretty infuriating because I do really believe X, but I’m obviously not doing violence about it. So it’s transparently false (to me) and the conversation is now about whether I’m being honest about my actual beliefs, which seems to undermine the purpose of having a conversation.
But it’s also kind of interesting, because it’s expressing a heuristic that seems valid—memes that incite “drastic” actions are dangerous, and activate a sort of epistemic immune system in the “uninfected.”
In this particular case, it’s an immune disorder. If you absorb the whole cluster of ideas, it actually doesn’t incite particularly drastic actions in most people, at least not to the point of unilateralist violence. But the epistemic immune system attacks it immediately because it resembles something more dangerous, and it is never absorbed.
(And yes, I know that the argument is in fact invalid and there exist some X that justify violence, but I don’t think that’s really the crux)
As an aside, I think a lot of “woke” ideas also undercut the basic assumptions required for a conversation to take place, which is maybe one reason the left at its ascendancy was more annoying to rationalists than the right. For instance, explaining anything you know that a woman may not know can be mansplaining, any statement from a white person can be overruled by a racial minority’s lived experience, etc., and the exchange of object-level information becomes very limited. In this case also, there is a real problem that wokeness is trying to solve, but the “solution” is too dangerous because it corrosively and steadily decays the hard-won machinery underlying productive conversation.
I think the underlying (simple) principle is that a pretty high level of basic trust is required to reach agreement. That level of trust does not exist on most of the internet (and I hate to see how much of the conversation on lesswrong now links to twitter/X, where it clearly does not exist).
As a “leftist” person, agreed! We’re all pretty fragmented. I’m working on creatively writing about these dynamics through the medium of a narrativized self. Sometimes the message doesn’t land as intended but all we can do is keep on trying to connect on the basis of shared consciousness (and faith in the goodness of everyone at their core).
“If you really believed X, you would do violence about it. Therefore, not X.”
I see this more often in areas other than AI risk, but regardless of context I agree it is a weak argument. It requires multiple other premises also be true. Let’s assume that people are not going to be moved by pacifism or violence is always wrong and just skip to practicals:
The violence one is capable of would cause a meaningful change in direction or time until onset of X
The targets of violence one has access to, are responsible for, or able to influence X
There will be no backlash to the violence that increases the likelihood of X
And these are just the basic ones that would rule out most reasons for violence.
If someone thought they could get away with eliminating a couple CEOs and data centers, would their belief in extinction risk justify doing it? Do they think it would change timelines? Because I don’t think it would. I give it below 1% that the loss of all the c-level execs at all frontier labs stops AI development. It jumbles up who gets there first, it changes who is personally taking what actions. But it doesn’t have a high chance of saving anyone, as far as I can tell. So bullet one applies. Also, I am quite confident that this would make the AI safety advocates lose ground, be portrayed badly in media (classic and social both), and thus it would more than 50% likely work against the goal of stopping AI risk (bullet 3), not for it.
There are very few people for whom violence is a realistic option. Because of these unstated premises.
The Gen Z riots globally , by some metrics have been doing pretty good. For example, the Wikipedia page claims 3 fails, 7 successes, and 14 ongoing with 4 of those alt least partly successful. This tallies to multiple governments overthrown, and several corrupt regimes having to back down on measures for self-enrichment. However, the new regimes appear to be just as corrupt, but favoring a different group. If anti-corruption was the goal, then bullet 1 applies and violence isn’t helping. If the goal is “the other team loses” then perhaps the Wikipedia claim holds up.
AI safety seems to be in a similar scenario where the most likely outcomes of violence are at best changing who gets their first, not whether we get there at all. I don’t see this as justification for violence, and to me the original claim is not valid.