The nostr protocol directly solves two problems in internet applications: identity portability and data portability. Products built on nostr can depend on those two problems being solved. For most other equally challenging problems, the nostr protocol has no single or opinionated solution, or lacks a publicly proposed solution entirely. This is part of nostr’s design, nostr's protocol “is just the way people do things.” For products that aim to address these other challenging problems, their solutions will exist outside of the protocol and most products will provide them in a centralized way. If these other problems are difficult, important, and numerous enough, your identity and data on nostr will be portable but many of your favorite experiences on nostr will be centralized.
When the wheel was invented, some people probably just saw a wheel - a stone that moves reasonably straight and unhindered once it starts moving. Others saw a wagon, fewer saw an automobile, and even fewer if any imagined a rocket. But the wheel is just a wheel. Given a wheel, if you wanted a wagon you needed an axle and if you wanted an automobile you needed an engine. Yet, after running “the way people do things” loop long enough, we have rockets.
If people want nostr to be more than portable data and identity (and of course, all the many impressive things that directly follow), if people want nostr to yield the decentralized social media equivalent of a rocket, if people want portable experiences that they can't be banned from or censored on, we should be trying to specify and decentralize the popular centralized experiences emerging on nostr.