The author didn't say it was fair, they said it was objective; objectively speaking, JS projects get all the new stars compared to backend and AI stuff, therefore are the most popular.
Of course, a more fair algorithm - if your definition of fair is that multiple disciplines and specializations, and / or language get the same exposure - would be to show the top starred / most active repos in certain categories (back-end, AI, specific languages, etc).
Isn't that super contradictory? How is the old algorithm both objective AND skewed towards Javascript?