Nothing surprising for someone who designs around this technology.
What Dave missed is that regarding g safety certifications, the choices are opening the breaker or limiting the output to below the breaker trip point (provided there are no touch temperature issues).
This is a huge advantage of PF corrected SMPS, output regulation, higher current efficiency and the ability to decrease the power supply voltage (or limit the audio) to decrease the AC current draw when needed.
Another thing Dave didn’t discuss is that at the high power level being discussed, the drivers won’t last much longer than the point that power fold-back occurs, therefore this is another way that increased speaker reliability is realized. In fact, most speaker processors do exactly the same thing with protective integrating limiter algorithms (where the threshold falls as the electrical analogy of VC temperature increases).
This limiting is not much of a thing at lower power such as bass amps. Also, 2 seconds of continuous tone at “clipping” (100% duty cycle) is not representative of any real world bass signal. It’s very much a fabricated situation.
Dave’s videos do generate a large number of views, and his presentation is quite click-worthy.