The logic is sound here, but I'm suspicious of the data on the topic and ability to make accurate assessments of these considerations. Mainly, I wonder about the population distribution getting sampled in IQ tests reflecting an entire race sufficiently, and reflecting possible subgroups. I've come across racial IQ literature links in convos like these before, but a lot of it seemed to come from a small sample. Like, compared to a modern clinical trial, the scale is way smaller, and even clinical trials end up missing various parts of the true population distribution in their sample. E.g., the city within a country tested probably has a gigantic influence. I'm inclined to believe this is a real barrier for drawing conclusions, because the numbers and SD given for groups on the bell curve seems to be highly variable from study to study.

On the topic, if we bring genetics into it, we wouldn't really be using race as something to connect to genetics. A scientific classification would probably be something like the mitochondria DNA of different population groups that yields tons of subgroups in race. There will be some overlap with population group associations and race associations, but it's really crude to lump all people of one color or continent together. This together implies there will be intraracial groups that can be exceptions, rather than only outlier individuals. That argues against racial generalization. Except in a case where the racial effect is extremely robust and strong, I suppose.

While I believe IQ measures something real that can be used to make predictions, that may just hold true at an individual level, given the limited data we have. Might be many other issues, given there are probable major confounding factors that co-associate with race. Frankly, I don't know real IQ tests enough to know whether they can compensate. Most of us in these conversations are only familiar with something like an IQ test you can get online and SATs (that are said to correlate well). Anyway, it would be nice if someone can address this point, but as I've heard IQ data does meet the best standards of psychology research, I'll ignore my concerns about the test for now.