@lostinseattle,
I perceive some healthy skepticism percolating below the surface which is perhaps why you found the article by Caelin interesting, but perhaps some thing gave you pause, which they should.
The article was written in a vacuum, and was not accurate in many regards. Was this intentional? I don't know. He claims extensive experience so if we take him at this word he was not truthful. Perhaps the untruthful statement is his experience.
Linear power supplies do draw power in peaks 120 or 100 times a second. Some switch mode power supplies do as well in addition to their switching. Some do not.
A false statement made was that restricting power\current delivery is bad. This is simply not the case. Almost every power supply intentionally restricts current delivery. I will state that again. Almost every power supply intentionally restricts current delivery. The ones that do not would probably fail EMI testing, and would have noisy outputs.
Big transformers in linear supplies are excellent current filters all on their own. Their inductance and the magnetic properties of the materials they are made with naturally filter high frequencies. That is often not enough, so high quality power supplies may add in choke (inductive) filters which provide additional high frequency filtering and soften current peaks. Filters are put around diodes to soften their turn on/off. Every switch mode supply has EMI inductors to soften the current spikes.
Caelin made a device that measures the high frequency performance of power cables. He conveniently ignores all the wire before the power cable will limit current delivery, as it must, but does not show, other than by making a false statement, that this correlates into better audio. One could argue based on his statements and knowledge of audio equipment, that it would make things worse.
