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.
"N years experience" loses value as a measuring stick when there are probably 2 orders of magnitude separating how much code the most passionate and least passionate developers wrote in that time frame.
The issues with diversity, prejudice and bias would be much easier if there weren't cases where something was both unjust and true. Software development is one of those activities where experience is helpful, so those who start earlier are generally better.
And it's not hard to guess why. Just because you've been programming for a long time, doesn't mean you were doing it correctly, or even understood what you were doing.
This "started coding in middle/high school or college" is a misleading metric since it assumes everyone has equal access and support to learn that way.
As some one from India who never had own computer until finishing college (the early 00's), along with no internet or cell phone, I find it hard to accept/respect those people who gloat about how early they started coding.
I always assume there are thousands of kids in the US who were/are not as lucky as the kids who started coding while they were in their diapers.
People/companies in US love to talk about their efforts to fix diversity, but without fixing this attitude of "earlier you learn, better you are" attitude, no amount of throwing money will fix it for the true diversity candidates.