It's a threshold filter. If you can create and maintain a project that's actually useful to people, you are definitely not a bozo. (NB: if you haven't done so, it doesn't mean you're a bozo; the implication only goes one way).

It's a better filter than "attended a super-elite university".

I think that it depends on what you're hiring for. This is analogous to Gladwell's, "there is no perfect pepsi, only perfect pepsis"

Ideally, githire should let one select from a few different hiring algorithms, which rank people differently.

Malcolm Gladwell is not a management consultant.

All I know is that if I put in Austin, Tx, I see a few people that I know in the top 10 to be really good software developers.

However, these are the types of people who generally start their own projects, and attract other good developers to a project. Their salaries are also going to be quite a bit higher than the average developer as well.

So, it seems to me that project popularity is at least some indicator that a person is competent as a software developer, but I can think of a lot of jobs that wouldn't require someone to be in the top 10.

I think that it would be great to be able to be able to search by code quality, test coverage, etc., but I don't think that we have the AI metrics to do that yet, and unfortunately, a real human has to look at someone's code to decide if it's any good.

However, this reminds me of software that tries to detect what "reading level" a person is writing at. For example, google lets you search by "reading level". It would be an interesting project for someone to apply the same principles to software.