Peering video content with a web seed fallback is actually a very sustainable and scalable solution. Content availability scales automatically with its popularity without increasing the cost to the host by using the upload bandwidth of users (which they generally have a surplus of).
The part that doesn't scale is not delivery (which can be parallelized independently among nodes), it is the custodial directory function -the tracking, location, identification and availability of high quality seeds and hosts. This is the function that will start out optimistically decentralized by hobbyists and enthusiasts and eventually succumb to opportunistic centralization as anything close to serious scale begins to kick in.
YouTube as a service has very different performance expectations than something like bittorrent. for one thing the directory function is built in and not externalized to another set of directory services, which are notoriously unreliable in the case of BitTorrent anyway.
And unlike BitTorrent or Mastodon I don't only want to know where I can find a given file, I want to know where I can find it delivered immediately, consistently, smoothly, without hesitation, in a variety of cross encoded formats I'm almost certain to have supported on my device, and with some sort of consensus mechanism to establish a better than chance confidence that the file I start streaming is the file I expected to stream and not something else with the same filename.
Nobody visits "the internet," though. People visit sites and services on the internet. People aren't going to go to PeerTube because it's decentralized. Viewers will go if there's content they want to watch, and creators will go if they can make money or they can have an audience.
Distributed and decentralized mean different things. distributed implies a distributor, which is itself a centralized function and in fact the internet has several of these centers of decentralisation that make sure that the grand partitions of high value 'land' assets such as IP address blocks and domain names are apportioned in an explicitly centralized manner.
'Decentralized' as it is commonly understood is about eliminating such central control points and devolving those powers to the edge of the network graph - to individual nodes, or more exactly to individual people.. This never works long-term, and always collapses right back into a centralized hierarchical system.
Information wants to be free, but value wants to be centralized.
The internet is so much more than a bunch of peering agreements. You can make a peering agreement with a network that you can't get transit to all you want, but if the underlying physical layer operator says no, you are out of luck. Maybe they'll let you in if you bribe them, maybe not. And this is Geography rules above all else here, which means concentration of access occurs where advantageous locations permit, and central points of control eventually monopolize these.
Fiber backbones and undersea cables and major internaps etc. cost a lot to build and maintain. The reason net neutrality is an issue at all is because the tradition of apolitical cooperative peering is under sustained attack from the natural forces of network centralization.
And you didn't address the issues I pointed out about the explicitly centralized 'land title' systems which distribute critical location identities to users. This is to say nothing of the ongoing tendency of higher layer de facto centralization of value location assets such as search and messaging.
Decentralisation isn't non-existent, it is just a temporary transition phase from one centralized hierarchy to another.
Yup, people who use the terms interchangeably tend to overlook to custodial and coordination problem solving paradoxes of decentralization and instead think of it only as a trendier-sounding variant of distribution.
It is best understood as a transitional phase between centralized regimes, not an end in and of itself.