In this case, saying damaging things some of which they expect will be shown to be false. He published after saying to NL that “I do expect you’ll be able to show a bunch of the things you said.”
I don’t think this means very much without specifying which things, and how relevant to the overall question they are. I would say this both in worlds where I expect something substantial or something minor to be inaccurate. I also don’t expect Ben would have said something different if he had waited longer for adversarial fact-checking. Reaching an epistemic status of “I don’t expect you’ll be able to show any of the things you said” seems extremely hard and unlikely.
I think it’s reasonable to criticize not waiting longer. But I object to using a sentence like this as really any evidence about the degree to which that was a mistake, or about the degree to which there are material errors in the post. Of course in any post like this, in a situation as adversarial like this, will there be some things that the post gets wrong. That’s true in any domain of this complexity. I think admitting to potential error helps in situations like this and interpreting things as adversarially as this undermines people’s ability to be honest and open to new evidence in situations like this.
I included Ben’s quote because it seems relevant, but I think I’d be saying pretty much the same thing without it (or if it turned out NL used Ben’s quote out of context). The “should have expected to be publishing some false claims just based on how many claims there were and how little time he allocated to checking the final claims with NL” is sufficient.
I don’t think this means very much without specifying which things, and how relevant to the overall question they are. I would say this both in worlds where I expect something substantial or something minor to be inaccurate. I also don’t expect Ben would have said something different if he had waited longer for adversarial fact-checking. Reaching an epistemic status of “I don’t expect you’ll be able to show any of the things you said” seems extremely hard and unlikely.
I think it’s reasonable to criticize not waiting longer. But I object to using a sentence like this as really any evidence about the degree to which that was a mistake, or about the degree to which there are material errors in the post. Of course in any post like this, in a situation as adversarial like this, will there be some things that the post gets wrong. That’s true in any domain of this complexity. I think admitting to potential error helps in situations like this and interpreting things as adversarially as this undermines people’s ability to be honest and open to new evidence in situations like this.
I included Ben’s quote because it seems relevant, but I think I’d be saying pretty much the same thing without it (or if it turned out NL used Ben’s quote out of context). The “should have expected to be publishing some false claims just based on how many claims there were and how little time he allocated to checking the final claims with NL” is sufficient.