If I try to imagine what happened with jessicata, what I get is this: taking responsibility means that you’re trying to apply your agency to everything; you’re clamping the variable of “do I consider this event as being within the domain of things I try to optimize” to “yes”. Even if you didn’t even think about X before X has already happened, doesn’t matter; you clamped the variable to yes. If you consider X as being within the domain of things you try to optimize, then it starts to make sense to ask whether you caused X. If you add in this “no excuses” thing, you’re saying: even if supposedly there was no way you could have possibly stopped X, it’s still your responsibility. This is just another instance of the variable being clamped; just because you supposedly couldn’t do anything, doesn’t make you not consider X as something that you’re applying your agency to. (This can be extremely helpful, which is why heroic responsibility has good features; it makes you broaden your search, go meta, look harder, think outside the box, etc., without excuses like “oh but it’s impossible, there’s nothing I can do”; and it makes you look in retrospect at what, in retrospect, you could have done, so that you can pre-retrospect in the future.)
If you’re applying your agency to X “as though you could affect it”, then you’re basically thinking of X as being determined in part by your actions. Yes, other stuff makes X happen, but one of the necessary conditions for X to happen is that you don’t personally prevent it. So every X is partly causally/agentially dependent on you, and so is partly your fault. You could have done more sooner.
If I try to imagine what happened with jessicata, what I get is this: taking responsibility means that you’re trying to apply your agency to everything; you’re clamping the variable of “do I consider this event as being within the domain of things I try to optimize” to “yes”. Even if you didn’t even think about X before X has already happened, doesn’t matter; you clamped the variable to yes. If you consider X as being within the domain of things you try to optimize, then it starts to make sense to ask whether you caused X. If you add in this “no excuses” thing, you’re saying: even if supposedly there was no way you could have possibly stopped X, it’s still your responsibility. This is just another instance of the variable being clamped; just because you supposedly couldn’t do anything, doesn’t make you not consider X as something that you’re applying your agency to. (This can be extremely helpful, which is why heroic responsibility has good features; it makes you broaden your search, go meta, look harder, think outside the box, etc., without excuses like “oh but it’s impossible, there’s nothing I can do”; and it makes you look in retrospect at what, in retrospect, you could have done, so that you can pre-retrospect in the future.)
If you’re applying your agency to X “as though you could affect it”, then you’re basically thinking of X as being determined in part by your actions. Yes, other stuff makes X happen, but one of the necessary conditions for X to happen is that you don’t personally prevent it. So every X is partly causally/agentially dependent on you, and so is partly your fault. You could have done more sooner.