Even if we had perfect filters to accurately identify the best talent, there's not enough of the top few percent to fill all the spaces in the industry, so someone is going to be hiring mediocre talent or forgoing having a business.
In the real world, though, we don't have perfect filters, and churn has a cost, too, so in practice most places are going to derive value if they can make effective use of mediocre talent rather than just letting it increase their churn.
(Moreover, one of the effects frequently claimed from great talent, employed effectively, as noted upthread, is not just their own direct output, but increasing the yield from lesser talent; if you don't hire any lesser talent in the first place, you can't benefit from that.)
Because you're solving mediocre problems for terrible pay. If you want to solve the world's toughest problems for terrible pay and can meet the bar, there are other places with a very happy customer, stable careers, and great benefits.
In my experience, that is what contractors are for; grind through the blergh B and C tier tickets. You can have onshore developers do that but a lot of companies seem fine with offshoring those tasks.
Because you have a mediocre job to do. If you hire excellent talent, you have to pay excellent prices. If you only have mediocre ROI tasks for said talent, then you're reducing the ROI of the task by overpaying for better talent.
Finding "the best" is very hard. You get such a small glimpse at a person from their resume and the interview process. Then good luck getting them to join you instead of another job that can pay more, give better perks, or whatever. You put in all that time interviewing, spending all that money, and look where you work, is it really all that effective?
If your company requires the best then you most likely have too much complexity. If your company requires the best to continue then your company isn't stable. Even if you got the best, can you keep the best? If you argue that there are enough "the best" then really you're just calling average (or anyone just above average) the best
The world isn't full of A players. Too many B players won't listen to grow even if they have the abilities. All you can do is help them discover their potential and that it's worth.
Because sometimes you know how to decompose a problem very well, and some problems are the nice kinds that have a nice horizontal scale-out for mediocre talent.