Economies of scale and network effects.
Nothing has to be the standard. But if 80% of people are familiar with X, choosing Y incurs friction.
For instance, there is an enormous benefit to the entire world from using a standard IP/TCP stack. Yet, we don't seem to suffer from a centralized authority making controversial or conflict-of-interest decisions regarding that stack.
So what I was suggesting is that document-editing abide by some global protocol/standard, which would include storage, versioning, and permissions. As far as the choice of user-facing interfaces and implementations of logic around those standards, there need not be just one, or even a few. Everyone can bring something cool to the table and anyone can use whichever flavor they wish.
It's a solution that works for anyone with basic IT skills. Offer your extended family access, and use the provided software to control permissions etc
For Docs for example: Sure, from a technical standpoint, creating a webapp that lets you modify a document, syncs over new changes on a regular basis, and lets multiple users collaborate on the same doc and the same time is certainly harder than making a hello world calculator, but it's definitely not that hard. Most of us could figure out how to do it, especially since it's 2021 and our hardware is up to the task.
However, what is stupidly non-easy, is getting billions of people to use your solution, and for it to become the "standard" in society, for powerusers and non-powerusers alike, and for it not to have hiccups when a billion people use it. Oh, also, you need to somehow pay for the physical infrastructure that supports it, and the human beings that maintain it (hardware + software).
The situation would be even more complicated for YouTube because of the nature of the content, i.e. a single video on YouTube might be viewed by billions of people, while even the most popular docs are viewed by thousands at most, and even then making a copy of a text document to then distribute off-platform is pretty trivial.
And yes, we can spit out terms like blockchain w/ proof of stake, IPFS, etc. - but the tech isn't even the hard part, it's everything else that's hard (adoption, consensus, complexity for non-powerusers, funding, etc.)