I hear it repeated like a mantra. A reassuring chorus, a superstitious incantation against the doubt that creeps in. “Relays are not stupid!” – spoken in turn, almost every day, in response to any anomaly, any suspicion of filtering, of shadow banning, of opaque curation.

It’s a phrase meant to calm the waters. For me, it sets off every alarm.

Because if those relays are not stupid – and they are not, they are code running on servers, code written by someone with resources, objectives, a vision – then they are doing something intelligent. And in an ecosystem that celebrates horizontality, neutrality, censorship resistance, “something intelligent” most often translates into an algorithm. An opaque, undocumented algorithm that decides what you will see, what you won’t, and in what order.

That phrase, “they are not stupid,” is the perfect cover for the most insidious of deviations. It transforms a potential act of active curation – of governance – into a mere technical fact, an “optimization.” It’s the language of the platform re-insinuating itself into the protocol. It’s the beginning of corrosion.

Deconstructing the Mantra

Let’s start from the principle: Nostr is fantastic precisely because it’s stupid. It’s a dumb protocol, in the noblest sense. A pipe. You throw a signed note into it, and it delivers it to whoever is subscribed to that pipe. Period. Its genius lies in this fundamental stupidity. It pushes intelligence, choice, filtering, curation to the edges: to the client, to the user.

But relays are evolving. They are growing. They have bandwidth costs, storage costs. Their operators see immense data flows, spam, potentially illegal content. And, being humans or organizations with finite resources, they introduce logic to manage it all. Up to this point, understandable.

The problem arises when these logics become:

1. Opaque: Not declared. You don’t know why certain content doesn’t arrive.
2. Asymmetric: Applied differently to different users.
3. Systemic: They are not a one-off decision on illegal content, but continuous filtering based on soft criteria (engagement, reputation, “quality”).

This is where decentralized enshittification is born. You don’t need a Zuckerberg at the top of a pyramid. A slow, imperceptible change of incentives at the critical node level is enough. The relay may not extract value by monetizing your data, but it can extract power or influence by becoming the de facto gatekeeper for a certain community. It can “clean up” the discussion according to its vision. It can, without meaning to, create a decentralized algorithmic bubble.

The Ghost Algorithm and the Death of Horizontality

When a relay introduces ranking, filtering for “quality” based on secret metrics, it is building a ghost algorithm. An algorithm whose parameters you cannot discuss, cannot audit, cannot consciously choose to use or not.

“But it’s their right! It’s their server!” Of course. Just as it was Twitter’s right, in the beginning, to be a small startup. The point is not legality. The point is fidelity to the spirit of the protocol and the long-term health of the ecosystem. Horizontality does not die with a coup, but with a thousand small cuts of convenience, optimization, “cleaning up the user experience.”

The opaque algorithm is the Trojan horse of surveillance capitalism in any environment, even a decentralized one. Because it replaces conscious choice, active search, the building of one’s own social graph, with a convenient, passive, curated feed. And curation, when it’s not yours, is always someone else’s project for your world.

Building Antibodies: Transparency, Choice, Roughness

So what to do? Surrender to the inevitable “maturation” (read: corruption) of the ecosystem? No. We can build antibodies.

1. Absolute Transparency as an Ethical Imperative: Every relay operated non-profit or communally should publish, on a .well-known page, its “Constitution.” What it filters and why. “We only filter demonstrable spam using these public heuristics.” “We do no ranking.” “We do not filter by keywords.” Transparency is the only disinfectant.

2. Clients that Expose, not Hide: Clients should have, front and center, an indicator: “These notes are coming to you from relay X, which applies the following documented filters: […]”. And let you compare the stream from two relays with one click. Make the invisible visible.

3. Celebrate Roughness: Resist the temptation to make Nostr “smooth” like traditional platforms. Roughness – occasional spam, the need to self-filter, noise – is not a bug, it’s a feature. It’s the sign that the choice is yours, not delegated. A healthy ecosystem is a bit wild, not a manicured park.

4. Protocol Incentives, Not Just Social Ones: Explore mechanisms at the protocol level that reward neutrality. A relay reputation system that certifies their “dumbness.” Tools for users to easily monitor if a relay is altering the flow.

Next time you hear “Relays are not stupid,” reply: “Exactly. And that’s why I demand to know exactly what intelligence they are using. Because on this opaque intelligence you cannot build a decentralized future. You can only build another walled garden, with more discreet gatekeepers.”

Nostr’s future is not in its ability to imitate the asphyxiating comfort of platforms, but in the courage to offer a more complex, more demanding, and infinitely more powerful freedom. A freedom that requires open eyes, sharp tools, and the refusal of any mantra that sounds like a lullaby for adults.

#NostrCritics #Algorithm #AskNostr #Decentralization #CensorshipResistance #Nostr #Moderation #Fediverse #Bitcoin #wotathon #FreeSpeech #OpenProtocol#NostrGrowth #NostrAdoption #WoT (Web of Trust) #NostrFeedback #NIP (Nostr Implementation Possibility) #NostrCritique