That's like saying you don't need search engines to view websites — your friends could just email you links to good ones. Technically true, sure.
Like, what you are saying is so nonsensical as to be not even wrong.
I /think/ you are trying to say something about discoverability, but like, it's not really that hard. You can just watch the local or federated tabs, and if you see anything interesting check it out. If someone seems alright you might even reply to something they post. Pretty soon you are having conversations, and the people you follow are boosting other nice, interesting people onto your home tab. Next thing you know, you are part of a community. It's great!
And all of this happens without the need for some weird corporate overlord to try and mediate the experience by way of manipulating you in order to keep you online longer so they can steal your data and shove ads down your optic nerves.
And where would the database live?
On the flip side, Micro.blog has a very different culture than Mastodon seems to[1], makes some very opinionated choices that many people might not agree with (e.g., there's not only nothing like retweets, but likes and even followers aren't public information), and of course, the easiest way to use it costs money.
[1]: Yes, I know there are lots of instances with different cultures, but there really is a kind of left-wing anarchist vibe across a lot of Mastodon; it's hard to explain if you haven't experienced it. Micro.blog comes across as generally more gentle, very conversational, and more Gen-X than millennial, if that makes any sense. (If it doesn't, sorry. Again, hard to explain if you haven't experienced it.)
Well, you used to be able to. They eventually closed it off, of course.
I think your concern is highly valid. Mastodon is a protocol-based network, much as SMTP email, NNTP Usenet, HTML+HTTP Web, IRC-based chat, and others are and have been. It's susceptible to the same types of challenges which have either centralised or killed off those platforms.
You mention sheer size. That's only one option.
There's the old Triple-E: Extend, Embrace, Extinguish. Applied successfully by Microsoft to DOS, the Office Suite, and Web browsers. By Google to Web Search and Web Browsers, to email, and increasingly to Web design and hosting. By Slack to Web Messaging. By Facebook to social media.
There's the risk of small instances fading away (happened to me with Mastodon.cafe) or large ones being bought (Mastodon.cloud). There's the prospect of admins going rogue (witches.town). Or, from the Diaspora* world, of admins dying with no succession plan in place (pluspora.com).
There are the Four Horsement of the Infopocalypse: drug-dealers, money-launderers, terrorists, and pedophiles. That's largely what killed off Usenet: the inability to defend against either malicious use of networks or those who made hay of that use to advocate for the shutting down of those networks. (This was combined with the lack of a compelling business argument for keeping them operating.)
There may simply be the staling of the protocol and paradigm. Mastodon is fun and interesting, yes, but lacks robust search, filtering, and privacy-management tools (e.g., auto-expiring posts), as well as obvious issues in scaling. If, say, a manageable instance size is 10,000 users, and Mastodon is to scale to serve an appreciable fraction of the roughly 10 billion people worldwide --- each instance administrator needs to at least consider 10^(10-5) or 10^6 other server relationships. It's one thing to keep track of reputations of a few thousand other servers, it's another to deal with millions, absent some robust tools for management, reputation, appeal, and similar factors. (My suspicion is that some degree of hierarchy with intermediate "hub" servers would emerge at some point.)
That said: Mastodon has proved robust and resilient beyond the predictions of numerous early critics. Not without flaws and issues, but those have to date been surmounted.