It being such a centralized organization definitely gives me "oh boy, here we go again" feels but it's not like it's not like typical users are going to choose anything better. Eschewing this is the essence of the perfect killing the good.
We don't need 'ultra decentralization' aka a different grocery store on every corner, nor do we want one single grocery store in the middle of the city, but rather just 'choice'.
Frankly it's like most things, we just want to avoid monopolization.
That said, a slightly different Mastodon configuration and a killer use case might enable it to rise aka business adoption.
As a designer and a long-time software developer, I see a TON of Dunning-Kruger Mt. Stupid comments about interface design, usability and what's reasonable to expect of users. Developers are naturally more willing to spend time futzing with social media technology because a) they have the domain expertise to reason about it comfortably, b) they have a high tolerance for high-touch software and take pride in their willingness to wrangle complicated and difficult interfaces, c) they care about the technical underpinnings, and d) because of all that, the space is likely to be populated by people who share their perspectives and interests.
If you take those things away, what incentive to users have to use the platform? I reckon that a lot of folks in the dev crowd don't naturally excel with the whole theory of mind thing, but if FOSS is going to produce tools useful to non-technologists, they should trust others that do. IMO the FOSS crowd desperately needs to embrace this cultural shift.
Not only do most users 'not care' about stuff, it actually is not that important ... until it is.
Twitter is actually 'good enough' for the most part. I don't like it, but I use it and find it acceptable. It's a big thing to ask people to shift gears for no really apparent reason (aka 'Musk says nasty things and is unfair!').
So yes, that is a primary product less: it has to be super, super easy to use and understand. Maybe, maybe, you can find a funnel to get people a bit more up the path of understanding. Maybe.
Which to me makes it just stupid of them to start from scratch rather than starting with interop with the Fediverse and optionally making breaking changes and/or bridging themselves.
E.g. the ability to migrate data more easily is great, but also easily accommodated by a combination of adding a new URI scheme and provide rewriting proxies to federate with ActivityPub instances that don't support it.
It feels like a lot of NIH and desire to be in control that I feel are at odds with building an open ecosystem.
In that spirit, I really enjoy:
1. The very good documentation this early in the project! This is not common.
2. The very rapid response developer team that does it live! Once the combination of surprisingly rapid membership growth + no blocks blew up and it became an urgent moderation feature, they had blocks within weeks despite the technically challenging task due to the kind of federated protocol and distributing blocklists. You can tell they have seasoned developers on the team. This is not just any kind of gimmick network trying to cash grab on Twitter exodus like I feel Hive Social was. It is an actual attempt at something better than Mastodon that has a fun, social experience with good onboarding and solving account migration headaches in mind.
3. The rare service disruptions despite the developers having their plates full and commonly introducing updates to the service. This speaks loudly about software architecture skills and being humble to risk management with good software hygiene. This again is not something that just happens but takes effort, experience and intent.
4. The exciting model of DNS approval which I still like. It's bloody fantastic to self-verify in a way that actually makes sense, and it feels very "World Wide Web" in a Tim Berners-Lee way. It uses pillars of the modern Internet in a way to strengthen a service and promises verification at scale. It can do company-wide verifications (domain.tld) as well as contributor-specific ones (username.team.company.tld). So, I dearly hope any misuse can be countered.
5. I worry they are overreaching with the AT protocol and federation but there is the "Shooting for the stars and aiming for the moon" saying here. I can only wish them the best and if I refer to the points above, the developer team seems surprisingly capable and full of actual intent here.