Replace goodwill with encryption. That's about data and metadata safety, but the same logic applies to everything else. No trust to people interested in breaking it.

As in - browsers' developers' goodwill was intended to keep Web standards' race in check. Protocols' extensibility was intended to allow for future backward-compatible development.

This was a wrong idea.

Gemini is one example of solving it, but one can imagine many others.

And it's fine if we have 12 Web protocols each for some specific idea of the Web. Among them some, say, would allow people to easily create webpages like year 1996, but sufficient for modern tasks and without all the bother with DNS and hosting (perhaps there is a p2p solution), Telegram shows that this is in huge demand. Many such variants are better than one overly complex, dangerous, corporate and oligopolized Web.

That's similar to how it seemed working anyway, we had e-news for global forums, webpages for personal pages, IRC for chats, ICQ\AIM\MSN for DMs, e-mail for reliable DMs, well, everyone remembers that time.

Nostr is a very raw, but maybe interesting idea for public social media.

Funny how Unix philosophy always shows itself in unexpected places, yes? =)