I actually disagree. It’s kind of a power law distribution thing where 1% of the people make 99% of the real scientific breakthroughs. We should absolutely optimize for geniuses.
That is maybe the single most dubious statistic I've seen thrown out so offhandedly in my entire time on HN.
Grants are reviewed and ranked by a panel of experts. While they often agree in the broad strokes, their scores also have random or subjective components: different reviewers might prefer different approaches; even the same reviewer might give slightly different scores if they re-reviewed a grant (e.g., due to their mood). However, someone eventually does need to set a payline, or threshold for funding.
As a result, you'd expect people who submitted the worst funded grants and the best unfunded grants, which have very similar scores, to have similar career trajectories. After all, their abilities--as measured by grant scores--are virtually indistinguishable. However, the winners go on to win nearly twice as much funding in the next eight years. This presumably reflects the fact that funding begets more funding (you can collect more and better data for subsequent work) and prestige begets more prestige (you've now got another thing to put in the "awards" category).
It's the Matthew effect multiplied by Goodhart's law.
https://en.m.wikipedia.org/wiki/Isaac_Barrow
and here
https://en.m.wikipedia.org/wiki/Fundamental_theorem_of_calcu...
Being smart certainly helps, but you also need the right opportunities, environment, and colleagues to “produce” science. These all feed back onto each other in complicated ways: funding gives you time and space to work on tough problems, which attracts talented colleagues, who can make you yourself smarter and thus able to attract more funding and more colleagues…and so on. This works the other way too: a funding squeeze limits what you can work on (if anything), etc.
I can certainly believe that it looks like a power law, but I doubt “intrinsic ability”, insofar as that’s even a thing, has a similar distribution.
Breakthroughs aren't made by lone geniuses anymore. We just have that superhero fetish eg. picture of black hole where credit was practically given to that one woman in the picture.
Bell labs was extremely successful (at least on the innovation front) not due to lone geniuses, but due to a large collection of very smart people all collaborating together.
I think the K12 system destroying potential geniuses is a seriously overblown concern, especially in the internet age.
The current system is terrible for genius kids' emotional and social development, and in some important ways it hobbles them. (Particularly when it comes to encouraging them to develop resiliency, accept that sometimes they need help, and develop their own sense of self-worth rather than derive it from the accolades of the adults around them).