He definitely states for each of his points that the people working on decentralized systems are misguided because decentralization doesn't automatically solve

He definitely states for each of his points that the people working on decentralized systems are misguided because decentralization doesn't automatically solve these hard problems; and while his thesis is what you state--that solving these problems is easier as a centralized system because you are more agile--other than a few awkward straw positions (many of which I addressed, and which aren't even really arguments but just anthropic assertions based on misstatements of history) he doesn't really have much to say on the idea that decentralization makes the problem harder. Hell: I'll admit he could have... I have myself given talks about issues with federated systems (vs. "truly distributed systems", which I think are very different than federated systems) and the pain of trying to deal with things like metadata privacy (something the people behind Matrix actually admit in some of their internal talks), but all of his positions here (with again, exceptions where he just ignores transport privacy or goes down rabbit holes of false tangents about phone numbers) really came down to a repetitive phrasing like "decentralized systems are not inherently encrypted, in fact most decentralized systems are not inherently encrypted by default; there's nothing about decentralization that makes things encrypted, you know?" (exact quote) that are later followed with a couple unrelated points and then a restatement of the (incorrectly argued) thesis that decentralized systems can't change and centralized systems are the only way to get new technology. It is like claiming "decentralized currencies don't automatically cause privacy or anonymity, and in fact the vast majority of cryptocurrencies provide more public records than your existing bank" while just flat out ignoring that things like Zcash not only could exist but in fact already exist.