There are diminishing returns and a ceiling on gains. The ceiling is quite high, but it’s there. You can only decrease disease risk to “roughly 0”, and this probably happens with medium-strength engineering (vaguely speaking); longevity probably (I speculate) has some caps coming from aging processes that aren’t fixed using genetic variants existing in the population; and IQ can’t be safely pushed way outside the human range. This means that countries could get ahead, but probably not crazy-ahead, and in the long run it should be easier to catch up than to pull ahead farther. (Ok this isn’t clear because maybe a head start on intelligence amplification snowballs or something.)
Uptake will start slow. It will speed up when people see how good the results are. But that information will be available to roughly everyone at roughly the same time: everyone sees the exceptionally healthy, capable kids at the same time, wherever those kids were. So for the big ramp-up, part that would matter on a national-national scale, there’s less of a head start. (There’s still probably serial lead time for some elements, but who knows.)
I do think there’s significant risk of inequality between ancestry groups, which relates to states though not one to one. That’s because there’s quite large inequalities between how much genomic data has been collected for different groups (see e.g. here: https://gwasdiversitymonitor.com/, though this is about GWASes, not exactly genomic data). Current PGSes don’t translate between groups very well. One way to address this is of course to gather more diverse data. (But the situation might not be so bad: plausibly once you more accurately identify which genetic variants are causal, your PGSes generalize between groups much better, or it takes much less additional data from the group to make scores that generalize.)
Other points:
There are diminishing returns and a ceiling on gains. The ceiling is quite high, but it’s there. You can only decrease disease risk to “roughly 0”, and this probably happens with medium-strength engineering (vaguely speaking); longevity probably (I speculate) has some caps coming from aging processes that aren’t fixed using genetic variants existing in the population; and IQ can’t be safely pushed way outside the human range. This means that countries could get ahead, but probably not crazy-ahead, and in the long run it should be easier to catch up than to pull ahead farther. (Ok this isn’t clear because maybe a head start on intelligence amplification snowballs or something.)
Uptake will start slow. It will speed up when people see how good the results are. But that information will be available to roughly everyone at roughly the same time: everyone sees the exceptionally healthy, capable kids at the same time, wherever those kids were. So for the big ramp-up, part that would matter on a national-national scale, there’s less of a head start. (There’s still probably serial lead time for some elements, but who knows.)
I do think there’s significant risk of inequality between ancestry groups, which relates to states though not one to one. That’s because there’s quite large inequalities between how much genomic data has been collected for different groups (see e.g. here: https://gwasdiversitymonitor.com/, though this is about GWASes, not exactly genomic data). Current PGSes don’t translate between groups very well. One way to address this is of course to gather more diverse data. (But the situation might not be so bad: plausibly once you more accurately identify which genetic variants are causal, your PGSes generalize between groups much better, or it takes much less additional data from the group to make scores that generalize.)