I don't exactly see how it clashes with heroic responsibility?

"When you do a fault analysis, there's no point in assigning fault to a part of the system you can't change afterward, it's like stepping off a cliff and blaming gravity."

Or what your skills are. People who are poor at soliciting the cooperation of others might begin to classify all actions which intend to change others' behavior as "blame" and thus doomed to fail, just because trying to change others' behavior doesn't usually succeed for them.

What could the woman in the harassment video do? Maybe she could start an entire organization dedicated to ending harassment, and then stay in NY as a way to signal she is refusing to let the harassers win. Or if the tradeoff isn't worth it to her personally, leave as Adams suggests. She isn't making it Scott Adam's problem, she's making it the problem of anybody who actually wants it to also be their problem. That's how cooperation works, and people can be good or bad at encouraging cooperation, in completely measurable ways. Assigning irremediable blame, or refusing to encourage change at all are both losing solutions.

I think taking responsibility for everything whether or not you caused in is exactly what heroic responsibility is about.

Apart from that Scotts get's a lot in the article wrong. In particular Scott argues:

The men in the street video are presumably repeat offenders. And that means they are getting a reward, at least occasionally, from their shouted public compliments. And I assume the reward comes from the occasional women who appreciate the compliments and smile back.

That's a naive view. It's probably wrong.

To the extend that Eliezer argues "Do the impossible" he doesn't argue doing things that literally have 0% of success. TDT discourages doing things with 0% of success. Eliezer doesn't argue virtue ethics where it matters that you try regardless of whether you succeed.

Not stopping with a naive view and actually working on the problem is something that Eliezer advocates and that's useful in cases like this. Even if it leads to questions that are even more politically incorrect then the ones Scott is asking.

I think I buy this line of reasoning in general, but I don't think Adams is applying it correctly in this case. If group A is doing something that makes you unhappy because group B is rewarding them for it, then it is no more "winner behavior" to go after group B than group A: in both cases you're trying to get others to fix your problems for you, by adding a negative incentive in one case and by removing a positive incentive in another.

I can make sense of this in a few ways: maybe Adams thinks at some level that B has agency as a group but A doesn't. (This is, clearly, wrong.) Or maybe he thinks that you're just more likely to convince members of B than members of A, which at least isn't obviously wrong but still requires assumptions not in evidence.