[zodiac]

The article is mixing two conceptually distinct points: that using github instead of CVs produces false negatives, and that it produces false positives.

It produces false negatives because lots of people who are good coders and who deserve to be hired don't have impressive github accounts for various legitimate reasons. I completely agree with this point.

However, to employers, false negatives are not that big of a problem. In some sense it's morally unfair that I exclude so many people who "deserve" the job, but if in the end I land a satisfactory candidate, I'm happy.

So to employers, the real question is, does github produce false positives? If my screening practice (I think we all agree that githubs and resumes should be used mainly as a screening step) is to ask for githubs instead of CVs, do I let better candidates through? If you read the article carefully, there is very little evidence presented for this; the author mostly argues that using github produces false negatives or weak signals.