There are also significant practical user experience differences.
Doing a search on Twitter searches Twitter, the whole thing. A search on Mastodon only knows about the servers you're connected to (unless you're searching for a specific user, then it'll micro-target their server to get their account info, but you have to know their name through some side-channel. Similarly, if you chance across a Mastodon post and want to follow that user, unless you happen to be on the same node as them you have to enter your own node data to get redirected to do the follow because of the domain-based nature of web security.
These aren't deal-breakers but we have the hard numbers from other web UX to know that every time you put a friction point like these in the flow, you immediately lose some x% of users. Relative to services that are centralized, these things will slow Mastodon adoption.
(This may not be the worst thing. There are other goals besides maximizing the adoption numbers.)