It’s not irrational to think that these range loses due to software updates are also being used in certain Model 3s as a proactive measure.
No, it's not irrational to think that. There may be some element of truth to that but I have not been tracking all the debate on that topic, so I really have no idea.
However, I think it is irrational to think that you will have 0% degradation over the course of your first year (or even half year!) of ownership. So if that's what you were seeing, and then it was followed by a sudden drop, it does suggest some sort of reporting issue resolved by software.
Regarding people driving down to 2-3% - maybe that's not low enough to see some sort of reporting issue. Many cars can be driven slowly for quite some distance below 0% (there's a couple kWh available). It may have been that this ability was compromised before the adjustments.
It's all speculation on my part here though; no idea really. Main point is just to expect degradation, mostly in the first year, and not the same for everyone.
(There was of course some initial speculation that Tesla had "solved" battery degradation with the new chemistry and cells in Model 3, but that does not appear to be the case. In some ways, the anecdotal data so far suggests it's actually worse than Model S. But I haven't seen an actual good analysis yet. But in any case, degradation is expected.)