In my view:
1. It is plausible that such treatments or screening would, at least initially, be more readily accessible to more affluent people, and thereby, being trans might become a stronger signal that someone is of low economic status than before.
2. A general sense that “this is a disappearing problem” might remove any incentive to work on reducing stigma faced by the still-existing trans people, providing them with better treatment, etc.
Sure, those problems are likely to be neutralized on a societal level by the fact that there will simply be less affected people left. But from the perspective of a single trans person, who obviously cannot benefit from PGT and might not be able to benefit from those treatments (e.g. if they only work during development), the personal net effect could easily be negative.
(To be clear, this is just my personal view and I haven’t yet had any other trans person articulate it to me exactly like that. There are of course also other objections to consider against PGT-style interventions to remove certain phenotypes from society, but I don’t think I have anything valuable to contribute there.)
That’s true, yes. Eradication is probably very hard, if not impossible. But I was more talking about speculative future screenings whereby you’d compute some “risk score” as a function of the full genome, select over that, and thereby merely reduce the number of trans people. Of course, it is very much possible that selecting embryos to minimize one kind of “risk” would just increase the rates of other types of problems and reduce genetic variance well beyond “reducing trans people”, so this kind of complex screening might not be worth it, either way, for any polygenic trait. It might not be possible to accurately predict the phenotypes of out-of-distribution genotypes at all, due to “computational irreducibility”-style dynamics. In the end, I don’t think that’s something that can be figured out on paper.
(Other traits for which stuff like this is being researched is autism and intelligence, and I think they serve as fairly good models. I think it’s fair to say that both are at least similarly complex, and if successes were shown with either, doing the same for gender dysphoria might be possible, too. Both show some highly rudimentary progress. Enough that I personally would assign at least a small probability that this is tractable.)
I don’t know if the point you’re making in your last paragraph is “this is unlikely to happen anyway, so as trans people, let’s not worry about it too much”, or “this is unlikely to happen anyway, so as a society, let’s reallocate efforts to more effective levers to reduce trans-related suffering”. Could you clarify? I think I would tentatively agree on both, but on the second point I do think it is worth asking whether it would be better to have less (or more?) trans people than we have at present, especially as such research is likely to become easier and cheaper as technology progresses in general. You know, just in case it does become possible.