He uses a metric for importance other than just size.
That said, his points that there were 600000 software developers in the US 10 years ago and more than a million GitHub members now, SO there must be significant developer mindshare seems nonsensical unless maybe if you assume GitHub is completely dominated by Americans, which anecdotally doesn't seem the case to me (I spent a few minutes looking for a breakdown of users by country, but didn't see any).
For sure about the confirmation bias, but I would propose that the post is meant to encourage scholars to come up with their own falsifiables and prove it one way or another.
The 600,000 vs 1,000,000 means: A lot of Americans, and the excess are people who are not from the USA or have multiple accounts. But the main point is that in gross numbers many of those 600,000 have accounts.
That said, his points that there were 600000 software developers in the US 10 years ago and more than a million GitHub members now, SO there must be significant developer mindshare seems nonsensical unless maybe if you assume GitHub is completely dominated by Americans, which anecdotally doesn't seem the case to me (I spent a few minutes looking for a breakdown of users by country, but didn't see any).
Also, confirmation bias.