I think we should recognize real differences in feasibility/difficulty/painfulness of actions and actionability of advice when they exist, for biological reasons or any reasons. (Sort of like how you wouldn't expect basic epistemic rationality advice to make someone with schizophrenia sane.)

We should also recognize the predictable effects of our words on people as they are, predicted using empathy and models based on people's actual behavior, rather than what we think people should be or non-truthseeking, habitually-used, constantly-surprised models of people. (Noticing when you're using the latter sort of models is a lot of work, but possible.) This might feel like abandoning all ideas of what people should be and letting them get away with any amount of laziness, and there are potential gains that could be lost that way, but the hard-ass approach loses at least as much (while making you less likable); far better to step back, recognize and (at least temporarily) let go of affective judgments and game-theoretic impulses, and semi-honestly try to figure out what's actually going on and what gains are possible.

The question I would ask is, does it help Alicorn to phrase your comment the way you did: "If you refuse to change your behavior in any substantial way I don't know why you're asking for advice." That would antagonize anyone, rationalist or not. If you said that to someone with schizophrenia, the last thing it would do is cure their disease. There are medications for that...and unfortunately, I don't think there are any medications for autism yet. And if anyone is bending backwards to accommodate it, it's Alicorn herself; this is something that must be extremely annoying on a day-to-day basis. You, on the other hand, don't have to change your day-to-day life at all.

That being said, I think your original suggestion (gradual habituation) was a good one. I don't know if Alicorn's tried exactly that strategy before, and there's a possibility it might work.