I should wait for the next parts, but I fear the author ignores our very important point of view: we are bayesian machines working with limited information under conditions of bounded rationality.

Lack of proof that Vitamin D supplementation is necessary is irrelevant. We don't work on proofs. If I were to describe our standard, it would a balance of "might it help?" and "might it hurt?". Currently, and I think taking the next parts in consideration as well, I think the answers will be very much in favor of supplementation - with caveats I'm actually very eager to read about.