He s talking about a mix of what Arthur called inward looking aliens, adaptation, and perfection of technology.
Three ideas that are not new, but maybe combining them in this way is insightful.
It doesn't answer the Fermi paradox. He basically answering "why don't we see Dyson spheres" with "because they are so inefficient that no one build s them." He does to a step further than that, to be fair, and promote an argument for why they are ineffecient. Why don't we see evidence of aliens? Because they do not want us to and have perfected their technology to be wasteless. But again... All of them? Th eres a problem with that, too. One of them is a fairly central tenant of modern physics, that no system is wasteless. We might be wrong, but as far as we know you're not allowed to just ignore entropy. And signs of that should be visible if there were a civilization that had harnessed an entire solar system. The nearer the approach to kardshev-3 the more undeniable.
but that argument existed about a day after Dyson supposed a swarm was possible. I mean at this point you might as well say "because they've enclosed our heliosphere in a giant holographic projector, because they know we're here and just want us to fuck off. Voyager will bump into it in 2143 and at that point the gig is up." Which i'm pretty sure is a plot in one of the Vernor Vinge novels. IF you can build a dyson sphere, if you allow that assumption, then a project like knowledge quarantine is less ambitious.
It's a fine argument and might be true. It's more of a philosophical argument though. I'm a fan of it in a different form, that the stable size of civilization is self limiting. Attainment in technological progress would be one factor among many, I would think. You don't get Kardashev 2 because while it is possible in principle, in reality it is selected against very strongly. So strongly that the nontrivial chances of it happening are on timescales longer than the projected lifespan of the universe... same reason all the air molecules don't just randomly pop over to one cubic foot volume in your room for a minute and then pop back to where they were. Kardashev is a very very 50's kind of idea. Or 90's. Remember Moores Law of computing and how we're supposed to have 50 godzillion hertz megacomputers in our wristwatches by now? That kind of thinking is necessary and wonderful, but it is only one mode. We could not have gotten to the moon without that kind of thinking.