I don't like this negativity. Pay and position doesn't necessarily reflect competence. I don't buy that the smartest engineers in the world are all working for big companies.
That's because people are obsessed with this idea of a one-size-fits-all great engineer.
A generally great engineer to a startup is somebody who can ship fast, make good enough decisions and trade offs, can think on their feet, and is a jack of all trades.
A generally great engineer to a big company is somebody who can work well with others, can follow process, can work within huge code bases and abstractions, and follow architectural best practices.
Google has brand recognition, and an enormous recruiting pipeline so they can hire one in 50 or 100 candidates.
They are better able to select for great candidates because they have more data and much more sophisticated system for identifying talent.
They pay very high salaries so they can retain top talent.
The average candidate that applies to Google is going to be a better than the average candidate that applies to a small/medium shop.
A small medium shop that tried this would be chronically and severely understaffed with a high turn over and would probably still fail at finding the best candidates.
If you are a small or medium sized shop and really need top level talent the best thing to do is probably just hire googlers and netflixers and pay them a market rate of 400-500k a year.
But for most products it's not worth it because the end product wouldn't be materially different enough from an average developer building it to justify the cost.
Also most small to medium shops could improve engineering quality by taking the Netflix approach which is getting better at firing people, and compensate them for the lack of job security with cash.
Optimizing for the rarest of scenarios - that you're going to hire the bestest engineers in all the worlds - is silly and foolish.
It's not negative. It's pragmatic. If you even happen to just encounter one of those people (they're rare by definition) then do what you can to hire them, but the truth is that the vast majority of companies producing software wouldn't even know what to do with them if they could attract them.
Stable, predictable, performance has a ton of value too - and you get a lot of that from 'regular' developers.
A generally great engineer to a startup is somebody who can ship fast, make good enough decisions and trade offs, can think on their feet, and is a jack of all trades.
A generally great engineer to a big company is somebody who can work well with others, can follow process, can work within huge code bases and abstractions, and follow architectural best practices.
These two do not necessarily overlap.