Have you considered whether your criticism itself may have been a damaging infohazard (e.g. in causing people to wrongly place trust in the CDC and thereby dying, in negatively reinforcing coronavirus model-building, in increasing the salience of the “infohazard” concept which can easily be used to illegitimately maintain a state of disinformation, in reinforcing authoritarianism in the US)? How many people did you consult before posting it? How carefully did you vet it?
If you don’t think the reasons I mentioned are good reasons to strongly vet it before posting, why not?
I have discussed the exact issue of public trust in institutions during pandemics with experts in this area repeatedly in the past.
There are risks in increasing the salience of infohazards, and I’ve talked about this point as well. The consensus in both the biosecurity world, and in EA in general, is that infohazards are underappreciated relative to the ideal, and should be made more salient. I’ve also discussed the issues with disinformation with experts in that area, and it’s very hard to claim that people in general are currently too trusting of government authority in the United States—and the application to LW specifically makes me think that people here are less inclined to trust government than the general public, though it’s probably more justifiable. But again, the protest isn’t about just die-hard lesswrongers reading the post, it’s about the risks.
But aside from that, I think there is no case to be made that the criticisms that I noted are off-base on the object-level are infohazards. Pointing out that the CDC isn’t in charge of the FDA’s decision, or pointing out that the CDC distributed tests *too quickly* and had an issue which they corrected hardly seems problematic.
The consensus in both the biosecurity world, and in EA in general, is that infohazards are underappreciated relative to the ideal, and should be made more salient.
Note that I pretty strongly disagree with this. I really wish people would talk less about infohazards, in particular when people talk about reputational risks. My sense is that a quite significant fraction of EAs share this assessment, so calling it consensus seems quite misleading.
I’ve also discussed the issues with disinformation with experts in that area, and it’s very hard to claim that people in general are currently too trusting of government authority in the United States
I also disagree with this. My sense is that on average people are far too trusting of government authority, and much less trust would probably improve things, though it obviously depends on the details of what kind of trust. Trust in the rule of law is very useful. Trust in the economic policies of the united states, or its ability to do long-term planning appears widespread and usually quite misplaced. I don’t think your position is unreasonable to hold, but calling its negation “very hard to claim” seems wrong to me, since again many people I think we both trust a good amount disagree with your position.
For point one, I agree that for reputation discussions, infohazards are probably overused, and I used it that way here. I should probably have been clearer about this in my own head, as I was incorrectly lumping infohazards together. In retrospect I regret bringing this up, rather than focusing on the fact that I think the post was misleading in a variety of ways on the object level.
For point two, I also think you are correct that there is not much consensus in some domains—when I say they are clearly not trusting enough, I should have explicitly (instead of implicitly) made my claim about public health. So in economics, governance, legislation, and other places, people are arguably too trusting overall—not obviously, but at least arguably. The other side is that most people who aren’t trusting of government in those areas are far too overconfident in crazy pet theories (gold standard, monarchy, restructuring courts, etc.) compared to what government espouses—just as they are in public health. So I’m skeptical of the argument that lower trust in general, or more assumptions that the government is generically probably screwing up in a given domain, would actually be helpful.
Cool, then I think we mostly agree on these points.
I do want to say that I am very grateful about your object-level contributions to this thread. I think we can probably get to a stage where we have a version of the top-level post that we are both happy with, at least in terms of its object-level claims.
Thanks for answering. It sounds like, while you have discussed general points with others, you have not vetted this particular criticism. Is there a reason you think a higher standard should be applied to the original post?
In large part, I think there needs to be a higher standard for the original post because it got so many things wrong. And at this point, I’ve discussed this specific post, and had my judgement confirmed three times by different people in this area who don’t want to be involved. But also see my response to Oliver below where I discuss where I think I was wrong.
Have you considered whether your criticism itself may have been a damaging infohazard (e.g. in causing people to wrongly place trust in the CDC and thereby dying, in negatively reinforcing coronavirus model-building, in increasing the salience of the “infohazard” concept which can easily be used to illegitimately maintain a state of disinformation, in reinforcing authoritarianism in the US)? How many people did you consult before posting it? How carefully did you vet it?
If you don’t think the reasons I mentioned are good reasons to strongly vet it before posting, why not?
I have discussed the exact issue of public trust in institutions during pandemics with experts in this area repeatedly in the past.
There are risks in increasing the salience of infohazards, and I’ve talked about this point as well. The consensus in both the biosecurity world, and in EA in general, is that infohazards are underappreciated relative to the ideal, and should be made more salient. I’ve also discussed the issues with disinformation with experts in that area, and it’s very hard to claim that people in general are currently too trusting of government authority in the United States—and the application to LW specifically makes me think that people here are less inclined to trust government than the general public, though it’s probably more justifiable. But again, the protest isn’t about just die-hard lesswrongers reading the post, it’s about the risks.
But aside from that, I think there is no case to be made that the criticisms that I noted are off-base on the object-level are infohazards. Pointing out that the CDC isn’t in charge of the FDA’s decision, or pointing out that the CDC distributed tests *too quickly* and had an issue which they corrected hardly seems problematic.
Note that I pretty strongly disagree with this. I really wish people would talk less about infohazards, in particular when people talk about reputational risks. My sense is that a quite significant fraction of EAs share this assessment, so calling it consensus seems quite misleading.
I also disagree with this. My sense is that on average people are far too trusting of government authority, and much less trust would probably improve things, though it obviously depends on the details of what kind of trust. Trust in the rule of law is very useful. Trust in the economic policies of the united states, or its ability to do long-term planning appears widespread and usually quite misplaced. I don’t think your position is unreasonable to hold, but calling its negation “very hard to claim” seems wrong to me, since again many people I think we both trust a good amount disagree with your position.
For point one, I agree that for reputation discussions, infohazards are probably overused, and I used it that way here. I should probably have been clearer about this in my own head, as I was incorrectly lumping infohazards together. In retrospect I regret bringing this up, rather than focusing on the fact that I think the post was misleading in a variety of ways on the object level.
For point two, I also think you are correct that there is not much consensus in some domains—when I say they are clearly not trusting enough, I should have explicitly (instead of implicitly) made my claim about public health. So in economics, governance, legislation, and other places, people are arguably too trusting overall—not obviously, but at least arguably. The other side is that most people who aren’t trusting of government in those areas are far too overconfident in crazy pet theories (gold standard, monarchy, restructuring courts, etc.) compared to what government espouses—just as they are in public health. So I’m skeptical of the argument that lower trust in general, or more assumptions that the government is generically probably screwing up in a given domain, would actually be helpful.
Cool, then I think we mostly agree on these points.
I do want to say that I am very grateful about your object-level contributions to this thread. I think we can probably get to a stage where we have a version of the top-level post that we are both happy with, at least in terms of its object-level claims.
Thanks for answering. It sounds like, while you have discussed general points with others, you have not vetted this particular criticism. Is there a reason you think a higher standard should be applied to the original post?
In large part, I think there needs to be a higher standard for the original post because it got so many things wrong. And at this point, I’ve discussed this specific post, and had my judgement confirmed three times by different people in this area who don’t want to be involved. But also see my response to Oliver below where I discuss where I think I was wrong.