Exactly: conditional on you already being a cyclist, it is much much safer to wear a helmet. Astoundingly, this fallacy accompanies every one of these "I quit helmets (and so should you!)" posts that I have ever read.
I think the argument is "I as a cyclist will be safer if there are more cyclists total because drivers will become more accustomed to them, and if I wear a helmet I foster the perception amongst other prospective cyclists that cycling is unsafe and they will thus be less likely to ride, so I'm better off trying to improve my safety via safety-in-numbers by projecting the image of cycling as a non-dangerous activity than I am in trying improving my safety via a helmet of questionable protective value." You can agree with that or not, but I don't think it's founded on a fundamental logical fallacy.
IOW, helmets are worthwhile as long as there as zero (or near-zero) marginal impact on cycling.
Of course, there is not zero marginal impact on cycling. Helmet advocates never seem to try to estimate what cost there is to cycling rates.
I suspect that even among committed cyclists, there is a marginal cost in the number and frequency of trips by bicycle to wearing a helmet, and this cost is much greater for more marginal cyclists.
You are making a point which is orthogonal to mine in that if you are conditioning on somebody cycling then they are by definition not on the margin. But no matter; the distinction is academic. I consider myself a "committed cyclist" and there is absolutely no cost to me of wearing a helmet. None. I feel naked and vulnerable without one, and these days pretty much won't ride sans casque, and that's the extent of it. (The thing about being a member of said group is, you've spent enough time in the saddle to see so many people doing so many stupid things on a bike, and a few of them pay dearly for it.)