Even the claims about him volunteering his own mistakes to Julia, which normally seems like a commendable action, could be interpreted as him trying to get ahead of any reports to a sympathetic ear inside the system. Given the community health team’s relative inaction about this case until the TIME article’s publication, it’s hard to discount this possibility without other information.
My impression is that many bad actors* are literally incapable of doing this. They never volunteer this sort of information. It’s bizarre – I don’t know why they don’t do it.** It would probably make them more credible if they occasionally disclosed extra information or admitted to having made mistakes you don’t already know about. In any case, because many bad actors are psychologically weird in this regard and seem incapable of admitting anything that makes them look worse than you already suspected, it still counts as evidence that someone is high on integrity if they proactively share information that had a good chance of not coming to light. I think this an instance where it makes sense, incentives-wise as well as epistemically, to give a person significant credit for disclosing things.
(Obviously don’t treat this as any sort of reliable test.)
*By “bad actors,” I mean people who don’t care about making others uncomfortable. I think it’s also bad if people repeatedly make others uncomfortable by mistake, but I want to treat these cases differently.
**One hypothesis is that maybe in the case of “bad actors,” the information that could potentially come to light would be so damning that they can’t risk admitting smaller mistakes that would have people look into these directions, where they might find out more. I don’t feel like this explains everything, though.
it still counts as evidence that someone is high on integrity if they proactively share information that had a good chance of not coming to light.
It might be reasonable for you to think that it wouldn’t have come to light but it would also be reasonable to think that it would. This really depends on how bad the case was, and how the interactions between OCB and the victim a few months ago went. Unless you have nonpublic information, we don’t know what the truth is here, and it feels weird to say that it “makes sense” to give him credit for this given these uncertainties.
Also, it’s possible that something has a good chance of not coming to light but that this still does not count as evidence that someone is high on integrity. For example: Owen sees Julia as someone who at least empirically has protected his interests. Or perhaps he thinks the case a few months ago was “less bad” than the masturbation case. Whatever the reason, he feels comfortable that sharing will be unlikely to lead to a bad outcome for him, and in fact may have a chance of reducing the risk of a bad outcome by getting in first with his version of events. If true, I don’t see why I should reward this behavior. The continued inaction of the community health team, and the fact the OCB did not suffer any meaningful negative consequences as a result of his actions prior to the TIME article could also support this hypothesis. I’m not claiming that this is what’s happening, but given this is a live possibility, I think you’re being too charitable here.
I agree that we should give no credit if there was a high chance of something further coming to light.
I guess I also agree that we don’t actually know what the status is of these other cases. (I don’t have non-public information.)
Owen’s account is here:
Was this incident an isolated case? Yes and no. I think this was by some way my most egregious mistake of this type. However, in my time in EA there have been four other occasions on which I expressed feelings of attraction towards someone in a way that — in retrospect as I’ve developed a more nuanced understanding of power dynamics — I regret. (In most of these cases I’m still on very good terms with the person.) I’ve slowly been improving my implicit models (so I never quite make the same mistake twice), but honestly it’s gone more slowly than I think it should have done.
I feel like saying “in most of these cases I’m still on very good terms with the person” isn’t the sort of thing that’s easy to lie about. If several future accusations were to come out, it’ll become clear that this was wrong, which would defeat the purpose of preemptively saying something.
Of course, for anyone who just hears the story itself and doesn’t have favorable priors about Owen from knowing him in other contexts, it makes sense for them to remain more cynical. (There’s also a bit of ambiguity about not being on very good terms with at least one person.) I approve of you pushing for that point!
One other thing is that, in my view, the way the apology is written provides some evidence that distinguishes whatever we want to call what Owen did from “common pattern of predatory sociopath.” (This is somewhat subjective, but I have another EA’s apology in mind from a situation that looks superficially similar, where the wording in that other apology was just so much more evasive and deliberately vague that this feels like night and day.)
There’s always the chance that my people judgment is wrong. I’m pretty sure I’m not the type of person who’s insufficiently cynical, but maybe I have a blind spot for a particular type of bad actor.
Was this incident an isolated case? Yes and no. I think this was by some way my most egregious mistake of this type.
Right-given his claim that this was by some way much less egregious than the TIME case, this should be an update that he would feel similarly confident that he’d take on very minimal risk from disclosing to Julia.
I can see where the predatory sociopath etc is coming from, but to be clear, all I am suggesting here is that just because something might have a good chance of not coming to light, this isn’t necessarily good evidence for a person being high integrity. The “coming to light” part is a proxy for “negative consequences”, so if Owen is sufficiently confident that there will be minimal risks of negative consequences of disclosure (indeed, because previous experience with a much more egregious case suggests this), and some indication that disclosing may be helpful for him in some way, this can be done in a self-interested way not indicative of high integrity. Importantly, doing this does not require Owen being a “predatory sociopath”, or ill intentions from him.
but I have another EA’s apology in mind from a situation that looks superficially similar, where the wording in that other apology was just so much more evasive and deliberately vague that this feels like night and day.
Just to be clear, you are not talking about Bostrom’s apology? I think I know what you’re referring to if not, and I’m not asking you to share it, but just thought I’d check as iirc we had a brief exchange about the Bostrom apology, and it would be useful for onlookers to know that you’re not referring to that if they don’t have context here.
“In most of these cases I’m still on very good terms with the person” is a hard statement to falsify. Unless Owen somehow pre-commits to who the individuals are in a way that could be revealed if necessary, we wouldn’t know if someone who came forward was one of the four. More importantly, it may be logistically and emotionally difficult for these four people to come forward in a way that protects their anonymity and allows us assurance that they are who they claim to be.
It takes just two people coming forward to falsify the statement instead of more than four, so the statement is more falsifiable than it could be.
(Another way in which it’s more falsifiable than it could be is because it makes claims about him still being on good terms with several people he has expressed interest in, which is falsifiable if the community health team or investigators were to ask him to mention the names of these people and then go talk to them. I’m not saying they need to do this – I’m just saying it’s good if people open themselves up to potential falsifiability.)
My experience is that people who habitually lie and deceive rarely constrain their options unnecessarily in this way. I want to emphasize that I see this as quite a strong pattern.”Keeping your cards hidden” is something I see really a lot in people I’d classify as bad actors. I’ve seen other apologies in similar context where this difference is like day and night.
This isn’t to say that signals like that can never be faked. It’s perfectly possible to hold both of the following views at once: (1) Some aspects of the apology are reassuring signals. (2) No matter how reassuring an apology is by itself, it makes sense to look into things a lot more, especially if there’s the possibility of a pattern.
My impression is that many bad actors* are literally incapable of doing this. They never volunteer this sort of information. It’s bizarre – I don’t know why they don’t do it.** It would probably make them more credible if they occasionally disclosed extra information or admitted to having made mistakes you don’t already know about. In any case, because many bad actors are psychologically weird in this regard and seem incapable of admitting anything that makes them look worse than you already suspected, it still counts as evidence that someone is high on integrity if they proactively share information that had a good chance of not coming to light. I think this an instance where it makes sense, incentives-wise as well as epistemically, to give a person significant credit for disclosing things.
(Obviously don’t treat this as any sort of reliable test.)
*By “bad actors,” I mean people who don’t care about making others uncomfortable. I think it’s also bad if people repeatedly make others uncomfortable by mistake, but I want to treat these cases differently.
**One hypothesis is that maybe in the case of “bad actors,” the information that could potentially come to light would be so damning that they can’t risk admitting smaller mistakes that would have people look into these directions, where they might find out more. I don’t feel like this explains everything, though.
It might be reasonable for you to think that it wouldn’t have come to light but it would also be reasonable to think that it would. This really depends on how bad the case was, and how the interactions between OCB and the victim a few months ago went. Unless you have nonpublic information, we don’t know what the truth is here, and it feels weird to say that it “makes sense” to give him credit for this given these uncertainties.
Also, it’s possible that something has a good chance of not coming to light but that this still does not count as evidence that someone is high on integrity. For example: Owen sees Julia as someone who at least empirically has protected his interests. Or perhaps he thinks the case a few months ago was “less bad” than the masturbation case. Whatever the reason, he feels comfortable that sharing will be unlikely to lead to a bad outcome for him, and in fact may have a chance of reducing the risk of a bad outcome by getting in first with his version of events. If true, I don’t see why I should reward this behavior. The continued inaction of the community health team, and the fact the OCB did not suffer any meaningful negative consequences as a result of his actions prior to the TIME article could also support this hypothesis. I’m not claiming that this is what’s happening, but given this is a live possibility, I think you’re being too charitable here.
I agree that we should give no credit if there was a high chance of something further coming to light.
I guess I also agree that we don’t actually know what the status is of these other cases. (I don’t have non-public information.)
Owen’s account is here:
I feel like saying “in most of these cases I’m still on very good terms with the person” isn’t the sort of thing that’s easy to lie about. If several future accusations were to come out, it’ll become clear that this was wrong, which would defeat the purpose of preemptively saying something.
Of course, for anyone who just hears the story itself and doesn’t have favorable priors about Owen from knowing him in other contexts, it makes sense for them to remain more cynical. (There’s also a bit of ambiguity about not being on very good terms with at least one person.) I approve of you pushing for that point!
One other thing is that, in my view, the way the apology is written provides some evidence that distinguishes whatever we want to call what Owen did from “common pattern of predatory sociopath.” (This is somewhat subjective, but I have another EA’s apology in mind from a situation that looks superficially similar, where the wording in that other apology was just so much more evasive and deliberately vague that this feels like night and day.)
There’s always the chance that my people judgment is wrong. I’m pretty sure I’m not the type of person who’s insufficiently cynical, but maybe I have a blind spot for a particular type of bad actor.
Right-given his claim that this was by some way much less egregious than the TIME case, this should be an update that he would feel similarly confident that he’d take on very minimal risk from disclosing to Julia.
I can see where the predatory sociopath etc is coming from, but to be clear, all I am suggesting here is that just because something might have a good chance of not coming to light, this isn’t necessarily good evidence for a person being high integrity. The “coming to light” part is a proxy for “negative consequences”, so if Owen is sufficiently confident that there will be minimal risks of negative consequences of disclosure (indeed, because previous experience with a much more egregious case suggests this), and some indication that disclosing may be helpful for him in some way, this can be done in a self-interested way not indicative of high integrity. Importantly, doing this does not require Owen being a “predatory sociopath”, or ill intentions from him.
Just to be clear, you are not talking about Bostrom’s apology? I think I know what you’re referring to if not, and I’m not asking you to share it, but just thought I’d check as iirc we had a brief exchange about the Bostrom apology, and it would be useful for onlookers to know that you’re not referring to that if they don’t have context here.
“In most of these cases I’m still on very good terms with the person” is a hard statement to falsify. Unless Owen somehow pre-commits to who the individuals are in a way that could be revealed if necessary, we wouldn’t know if someone who came forward was one of the four. More importantly, it may be logistically and emotionally difficult for these four people to come forward in a way that protects their anonymity and allows us assurance that they are who they claim to be.
It takes just two people coming forward to falsify the statement instead of more than four, so the statement is more falsifiable than it could be.
(Another way in which it’s more falsifiable than it could be is because it makes claims about him still being on good terms with several people he has expressed interest in, which is falsifiable if the community health team or investigators were to ask him to mention the names of these people and then go talk to them. I’m not saying they need to do this – I’m just saying it’s good if people open themselves up to potential falsifiability.)
My experience is that people who habitually lie and deceive rarely constrain their options unnecessarily in this way. I want to emphasize that I see this as quite a strong pattern.”Keeping your cards hidden” is something I see really a lot in people I’d classify as bad actors. I’ve seen other apologies in similar context where this difference is like day and night.
This isn’t to say that signals like that can never be faked. It’s perfectly possible to hold both of the following views at once: (1) Some aspects of the apology are reassuring signals. (2) No matter how reassuring an apology is by itself, it makes sense to look into things a lot more, especially if there’s the possibility of a pattern.