I think you're equivocating on "past selves" between, "A decision I made four months ago when my head was clear", and, "What I did five years back and I've learned a lot since then".

Or, more formally, assuming continuous self-change, I think the quote you cited is talking about a period of time long enough for the amount of self-change to be sufficient that the person you were then is different to the person you are now.

I reject the premise that that period of time is quantifiable independently of the proposition in question. It isn't a question of whether I'm a different person in general; it's (ideally) a question of whether anything has changed in the premises on which I based my thoughts about the particular situation. One decision I considered ten years ago may still be valid if the situation comes up now, while another that I considered last week may already need revisiting.

Oh, sure, I completely agree with that.

But yeah, deciding in any given case whether the virtue of self-trust or that of self-distrust is the right one to apply is certainly the fun question :-)

Exactly. Except that I would have thought that a lot of the time it in fact it ought to be reasonably clear which applies.

The quoted sentiment seems to be "when you realise you used to think something different [and implicitly: that the new thought is, on considering both, obviously the correct one] you shouldn't attempt to justify the old thought, but to accept that you changed".

Conversely, the situation you describe is I think generally when a temporary emotional response overrides a previously clear decision (and implicitly that the previous decision was likely to be correct).

However, once you get as far as diffing the old-you and new-you attitudes, it seems likely that which is appropriate will be obvious.