Okay, great! I'll assume some baseline competence in the area, but still write for a somewhat general audience because others may be reading.

I've read The Bell Curve too (roughly when it came out). It was interesting, but mostly not very impressive in terms of its persuasiveness because the core problem was that basically every measure of IQ confounded social and genetic factors. I don't remember the details (I am not sure whether I kept the book in storage or not), but what occurred to me is that you cannot make the same flawed measurement in a bunch of different contexts and conclude that the thesis must be true. The problem is that every flaw is of the same type and in the same direction.

In the racially diverse countries examined (e.g. the U.S.), race is confounded with culture and economic attainment, and in the racially homogenous countries examined (e.g. countries in Africa vs Scadinavia), race is confounded with both culture and economic attainment.

And it's not the case that those things don't matter, because you can note the correlations between economic status and IQ among people of one race, and you can note that because of the Flynn effect (https://en.wikipedia.org/wiki/Flynn_effect), the IQ in the United States a hundred years ago was lower than is the IQ in Africa now (and the selective pressure is insufficiently strong for there to be any mechanistic way for the Flynn effect to be genetic--indeed, recently, the selective pressure seems to be weakly in the opposite direction). So clearly environment plays a huge role (how huge depends on context, apparently), and the environment of people of different "races" is...well...different. You can't ignore it.

Another way to go would be to identify polymorphisms associated with intelligence differences, and check both whether the effects were similar in people of different geographic origin, and if yes, then count up the prevalence of favorable alleles in different populations. Unfortunately, our best efforts to identify the actual alleles (not just some nebulous "it is genetic") have been underwhelming. For instance, https://www.nature.com/articles/s41588-018-0152-6 reports about 5% of variance attributable to the alleles it's identified. So that's right out. You can't make any meaningful conclusions when 95% of the effect is caused by factors you haven't identified.

So we can't do a fair direct comparison, and we also can't do a competent genetic comparison. This leaves us with correlational observational studies. These are problematic in the way observational studies always are--it's hard to tell which variables are causal and which are not, and usually people only do covariate analysis rather than something more complex like structural causal modeling.

But it's the best we've got, and what tends to happen is that, almost like magic, once you start plugging covariates in, the racial gap dwindles and dwindles and dwindles until it's lost in the noise, and even if the small residual is real, it's completely dwarfed by individual variation.

For instance, LeWinn et al., JAMA Pediatrics 171:1-14 (2020) show that after correcting for 24 significantly correlated variables, few of which seem like they could be affected by genetics, the remaining explanatory effect of race has a 95% confidence interval of -0.14 to 2.26 (advantage of categorical variable of race as "white" instead of "black"). (Note--you have to look at eTable 3 in the supplementary material to get these numbers.)

Now it's possible that some of the covariates are actually determined by race. For instance, parental education level is one factor that matters, so if there were a race-based difference in IQ, then that would show up in less educational attainment, which would be incorrectly assigned to education even though race was actually causal. So it might be an underestimate.

But it's also possible (quite likely, actually) that other correlated variables exist that we haven't found yet, so the difference might be an overestimate. For instance, there is no correction for other aspects of culture (like what attributes in children are typically praised by adults), but for historical reasons there are on average cultural differences between black and white people.

Regardless, because we know (because we can intervene and see considerable differences) that going to grad school and such is impacted substantially by various cultural factors, even if there is some correlation in that regard, it's unlikely to be the dominant factor for black people.

So on the face of it, the best-in-class research in the area says that there's basically no difference (maybe a couple of IQ points) on average between "black" and "white" people. And when thinking more deeply about it, it seems like even that small and not-statistically-significant difference is likely to be an overestimate of the difference, and even if it's an underestimate it wouldn't be underestimated by much.

Thus, the best evidence is: any genetic difference between blacks and whites in the U.S. is small to zero, far below individual variability, and basically is a non-issue.

Now it is true that intelligence ("g" especially) is a real thing, and that it is highly heritable. However, like with most traits that vary in humans, where the range of variability within a group is large compared to the mean difference across groups (and this is true with most allelic variation as well), intelligence seems to be mostly an individual trait, not a group trait.

I'm not going to share any stories about being smart or not; I think the relevant thing here is the quality of the thought process that has gone into this message, and that is, I trust, clearly on display.

But I will point out that when I find good evidence for a point of view that conflicts with what I previously believed, I have a strong tendency to update my beliefs. This is, I think, a mark of intelligence.