Although most markets have historically had a high concentration of output from the top percentile, so its not exactly a far fetched proposition to say that 5% of the workforce led development during the IR.
If the vast bulk of technological output amplifications are additive, then we'd expect to need broad distribution of the skills and experience a broad distribution of results.
If even a few technological output amplifications are multiplicative, and especially if they are independent (one can freely choose the 3x and the 4x and the 5x advantage and obtain something like a 60x advantage), then we would expect the skills to end up concentrated and the optimal strategy to be to load up as many multiplicative advantages as possible in one place.
(Note carefully the first few words of each paragraph. It only takes a handful of multiplicative advantages for them to dominate.)
You can obtain this result by playing lots of strategy video games, especially Civilization which has both. Additive results like the simple Granary, which adds a constant amount of resources to the containing town, need to be built in every city to be effective. But things that have a multiplicative advantage need only be built where they are necessary, and if you have certain unique things that can only be built in one place that are also multiplicative, the optimal strategy is to build them all in one place so they reinforce each other, not scatter them about.
So far, none of this has a moral dimension. This is all just simple optimization. Now we bring it back down to reality, where morality intrudes. It should be so obvious that a great deal of technology in the real world has a multiplicative effect rather than an additive effect that I should not have to justify that statement. IMHO, the question is what to do about this rather than whether this is true. And one of the dangers is that it is very easy to end up giving answers that actually destroy the technological advantage in the process, because for as mighty as it all appears to be, it's a lot more fragile than it looks, as you can see once you start seeing it this way. The humans involved in the tech aren't irreplaceable (or it would be really fragile), but there's still limits involved, and destroying or impeding their effectiveness can have non-linear impacts on total productivity. A lot of people model the technological advantages we've obtained for ourselves as a constant; we can do whatever we want and it will always be there, no matter how we impede or help it. It's scary how untrue that is.
Although most markets have historically had a high concentration of output from the top percentile, so its not exactly a far fetched proposition to say that 5% of the workforce led development during the IR.