The name “The Human Web” is doing double duty.

It's the World Wide Web, obviously. But it's also a web of trust: people connected to people through deliberate choices about who to follow, who to vouch for, who to amplify. That second meaning is the one that matters.

The web was always social

The web was social before Social Media. Blogrolls, RSS feeds, forum signatures linking to personal sites, people vouching for people. If you found a writer you liked, you checked who they linked to. Their blogroll was a recommendation: these are voices I think are worth your time. You followed the links, made your own judgment, and maybe added a few to your own feed reader. The chain was visible. You could trace how you found someone, and you could walk away whenever you wanted.

That was a social structure. It was organic, messy, and human-scale. It wasn't optimized. Nobody was A/B testing your blogroll to maximize engagement. The signal was low-volume but high-trust, because every connection was a choice someone made.

Social Media didn't add the social layer. It replaced an organic one with an algorithmic one. “Who do you trust?” became “what gets engagement?” The connections between people stopped being the mechanism by which content reached you. The algorithm took over that job, and it doesn't know or care about trust.

What a web of trust looks like

You follow someone because you've read their work and decided they're worth your attention. Maybe you found them through the firehose, maybe a friend recommended them, maybe a platform surfaced them based on your existing trust graph. The discovery mechanism doesn't matter. What matters is that you made the choice. They follow others. You don't know those people, but you trust the person who trusts them. That's a web.

It's how recommendations have always worked offline. Word of mouth. Referrals. “My friend knows a guy.” You don't trust the guy directly. You trust your friend's judgment, and that's enough to give the guy a shot.

Trust has distance. First-degree connections are people you follow directly. You chose them. Second-degree connections are people they follow, one step removed. Beyond that is the open firehose: everything, from everyone, unfiltered.

You can always explore the firehose — nobody is stopping you — but you know when you're doing it. The web has edges, and you can see them. That visibility is the point. Not restricting what you can see, but making sure you always know where you stand.

Following is complicated

Here's where it gets honest.

Following someone can mean very different things depending on context. Sometimes “I follow this person” means “I vouch for their judgment.” Their recommendations carry weight with me. If they trust someone, that's a meaningful signal. Other times it just means “I want to see their stuff.” Pure interest. No endorsement implied.

These often overlap, but not always. You might trust your accountant's judgment completely and have zero interest in reading his essays. You might enjoy a provocative writer without trusting their judgment enough to let them shape your trust graph.

And the balance between trust and taste depends on what you're consuming.

Consider longform writing. If you respect someone's thinking enough to read their essays, there's a decent chance you'd find value in their reading list too. Following is a strong signal. It says something about judgment, not just preference. A recommendation from someone you follow carries real weight.

Now consider music. You have diverse tastes. Punk, orchestral, folk, whatever. Who you follow on a music platform means almost nothing to anyone except people who happen to share your exact preferences, and even then maybe only in one genre. “Trust” barely applies. It's almost purely taste, and taste is multi-dimensional. Your follows are a scatter plot, not a signal.

The correlation of trust and interest, then, is a spectrum. The more a medium rewards sustained judgment, the stronger the trust signal when someone follows. The more it rewards visceral, aesthetic preference, the weaker. Longform writing sits at one end. Music sits at the other. Film, podcasts, short video, they all land somewhere in between.

What this means for The Human Web

This is where a lesser project would try to paper over the complexity. Define “follow” precisely. Build one trust algorithm. Ship it.

That would be a mistake. Baking a specific trust-propagation model into the infrastructure would make it wrong for most applications. A model tuned for longform writing would be meaningless for music. A model tuned for music would be useless for journalism. The spectrum proves that no single answer works.

So The Human Web doesn't try to answer it.

The Human Web defines three things universally:

Identity. A keypair. A signature. Continuity over time. The foundation from the crypto primer, unchanged.

The follow graph. Who follows whom. A portable, verifiable record of the choices people have made about who to pay attention to.

The principle. Trust propagates through human choices, not algorithms. What you see is shaped by people you chose to trust, not by an engagement-maximizing black box.

What The Human Web does not define: what “follow” means in your application. How trust propagates through the graph. How content gets filtered or surfaced. That's yours to build.

This is by design, not a cop-out. Roads don't define whether you're driving to work or to the beach. They give you the infrastructure to get where you're going, and where you're going is up to you. The Human Web works the same way. It provides the foundation. Applications build meaning on top.

A longform writing platform can treat follows as strong endorsements and surface second-degree content aggressively. A music platform can treat follows as loose interest signals and rely on other mechanisms entirely. Both are building on the same identity layer, the same follow graph, the same principle that humans decide what's trustworthy. They just interpret the data differently, because they should.

The old social media

There's a recurring theme in these essays: The Human Web isn't proposing something new. It's proposing a return to something that worked, hardened with cryptography.

The blogroll era had the right structure. People vouching for people. Organic discovery through human choices. Legible chains of trust. It lacked the tools to scale it, verify it, or make it portable across platforms. The Human Web adds those tools.

The new social media looks a lot like the old social media. People choosing who to follow. Recommendations flowing through human relationships. The ability to see where content came from and why it reached you.

What it doesn't look like is Social Media: the algorithmic feed, the engagement optimization, the opacity, the platform as gatekeeper of identity and attention. That's the thing being replaced, not the social web that came before it.

Trust requires judgment

One thing worth saying plainly: a web of trust is only as good as the humans in it.

Some people will build echo chambers. Some will follow exclusively within their own tribe and never look at the firehose. That's their right, and at least they're doing it to themselves with their eyes open, rather than having an algorithm do it to them in the dark.

Some people will be bad judges of character. They'll vouch for someone who turns out to be a fraud. That's fine. In an algorithmic feed, the damage from a fraud spreads to millions before anyone notices. In a web of trust, you prune one connection and move on. The blast radius is limited by the structure itself.

The web of trust doesn't eliminate bad judgment. It makes bad judgment visible, local, and recoverable. That's a significant upgrade over the alternative.

It's not a platform

When I say “The Human Web,” people hear “a new social media platform.” It's not.

Your identity is a keypair. Your follow graph is a set of signed declarations. These are yours. They're portable. They work on any platform that honors the protocol. The Human Web is infrastructure, not a destination.

Different platforms will build different experiences on top of it. Some will be for writing, some for music, some for conversation, some for things nobody has thought of yet. They'll interpret follows differently, surface content differently, define trust differently. That's not fragmentation. That's the system working as intended.

The web of trust is the connective tissue. Your identity and your relationships travel with you. The applications are just lenses.