This mindset seems to work rather well for more objective areas like athletics. I don't see any reason why an intellectual pursuit like programming would be any different.
But at least, in athletics, there's no charade of trying to fix the gender/race bias by throwing money at programs that make other's catch up so that they can claim "equality".
Most Marathoners and NBA players often are African or have African ancestry. And no one says the same team should have an equal number of <insert race/gender here>.
IMO, humans are adept at learning intellectually for a much longer time, compared to improving physically in 'objective areas' like athletics, which needs to begin very early in a human's life.
1. Athletic performance is easy to measure. Especially for individual sports (track and field, gymnastics, weightlifting) the numbers are the only way to compare athletes. Not so for programming. If someone invented a way to definitively judge programming ability today they'd become very rich, very soon.
2. If we assume that mastery of programming follows a sigmoid-curve (not an unreasonable assumption, IMO) it means that the ability difference in "coding for X years" and "coding for X + 5 years" becomes nearly indistinguishable for some value of X, for people of roughly similar innate ability and drive to improve. And if you believe in innate ability, then there will certainly be people who came to the field late and caught up with, or surpassed, less-innately-talented colleagues who have been programming longer.
3. One of the reasons that starting at an early age is important in athletics is because peak performance is harder and harder after you turn 30. The "growth" phase has to be therefore significantly accelerated. There's also evidence that for technical sports requiring motor coordination (eg. soccer), learning young is better. Whereas people can code at peak level until the day they die, assuming no other illnesses.
Architects do their best work very late in life.
I personally wouldn't want a surgeon that started cutting into people at age 9.
See anyone you can think of from the 60s and 70s like Ken Thompson and Dennis Ritchie.
Ken Thompson had a MS in EE/CS and Dennis Ritchie had degrees in both physics and mathematics, all of which are far more difficult than programming. Most of the other great names of the '60s/'70s have similar backgrounds, invalidating your point.