I think what he's talking about (not bothering to re-read) is the power required to magnetize the alternator components in the first place, so the alternator works at all. They use electromagnets rather than permanent magnets, because electromagnets are much cheaper, especially at that size. This is a fixed power requirement, so if the alternator can't generate more than that fixed power need to operate, there's no point in running.

Magnetizing power (and other constant-term losses) are a very small portion of the rated turbine power in a doubly-fed induction generator (DFIG).

Edit (one more thing): At this scale, permanent magnet (PM) machines aren't much more efficient than DFIG machines. It turns out that eddy current losses in the stator back iron remain as a significant term in both machines. But for a DFIG machine, the eddy current losses at light load are reduced by quite a bit relative to the losses produced by a PM machine. Since real-world wind turbines end up operating at partial load for much of their life, the effects basically wash out, or are even a net negative for the PM machine.

As an aside, the use of electromagnets occasionally produces hilarious failures. There have been some rare signaling errors that push a huge amount of power through the alternator and turn windmills into giant fans (at least until they shred themselves).