On what the ‘stop’ is doing: my impression is that it’s fundamentally arguing from a position of assumed shared values.
If someone were to split me off into two initially-identically-similar copies who then went our separate ways for a time, diverged thereby, but eventually returned to engage in some cooperative venture, and the other me suddenly stared at me intensely and told me to stop in the right sort of forceful tone, I would be likely to stop what I was doing. I share values with myself; if she tells me to stop, in a manner conveying urgency, and doesn’t explain further, I assume that this is because there’s a good reason for me to stop; perhaps I’m about to mix some chemicals which, unbeknownst to me, will explode on contact. She’s not an authority above me, but she’s an authority coextensive with me, able to order me around for much the same reasons that I’m able to order myself around.
However, my likelihood of remaining stopped will decrease with time if no explanation is forthcoming; after all, while “those chemicals are about to explode!” takes longer to say than “stop!” does, it doesn’t ultimately take that long. The longer it takes for such an explanation to come forth—especially if the moment appears otherwise unheated, as opposed to being followed immediately by commands like “put that bottle down gently and then run!”—the likelier alternative hypotheses become, such as that actually the other me was mistaken about whether stopping was the correct action for me to take, or that she’s exploiting my trust in our shared values to convince me to take some action which in fact is good only for her and not for me, or suchlike.
This explains why it feels reasonable as an action to take, in the face of the cruelty discussed in this post: large fractions of people share enough values, and have enough awareness of this value-sharing, to be willing to stop-when-told-in-the-right-manner on at least a short-term basis. It also explains why there’s a range of plausible slightly-longer-term outcomes which includes both “people embarrassedly shrink down and walk away” and “people laugh and get right back to attacking the person” following the stopping: some will take the pause as a chance to reflect and notice that stopping is in fact good for them, while others will reflect and conclude that your telling-them-to-stop isn’t actually good evidence that they should do so.
(I have less in the way of confident explanation for the sense of Godly Anger. Perhaps I would have more if I were disposed towards feeling such things myself; but, for the most part, my instincts about the described scenario are the straightforwardly-subjectivist “these people are doing a thing I consider bad, and I accordingly want to stop them”, with my sense that this might be echoed on a more global scale being only instrumental and tactical—factoring into my estimates of how likely they are to be stoppable via verbal command, how likely I am to be able to find allies in stopping them by force, et cetera—rather than morality-flavored.)
One question which this interview touches on a bit but doesn’t address directly, which I find myself curious about: what were the core elements of the story which were The Point, which the rest of the story got built around? Which is to say: some parts of the story, like the SPHEW arc, were added as setup for later parts which needed to be there. But it doesn’t make sense to write something for the sake of setting up something else, if the regress doesn’t end with something you’ve outlined for reasons other than setting up something else. The terminal story-elements, so to speak, as opposed to instrumental ones. I’d be interested to know which elements of the story were, in the relevant sense, terminal.