Thread

Pendulums

I’ve been thinking about the internet a lot these days. How it started, how it evolved, and where we’re at today with social networks. It appears as though we’ve always had this constant battle between decentralization and centralization; a pendulum swinging back and forth over the last 30-40 years. When I think about where we are with social networks, it’s a really odd form of centralization. It’s our choice to be here, why are we? I do not believe they’ve centralized all content (websites still exist). I do not necessarily believe it’s to keep up with our friends (“For You” feed usage vs “Friends/Following” feed usage). I believe it’s because they’ve succeeded at centralizing attention.

Algorithm-driven communities

Because our attention is centralized, some refer to social networks as “digital town squares” – are they? Town squares are the social, political, and economic heart of a community, and unless you’ve interacted with the algorithms long enough to fine-tune the feed, are you in a digital community? A community is a group of people who share something in common; places, interests, identity, values, or purpose. Aside from the same app (place), what are they without algorithms? Without algorithms, are these places you return to, to see the same people? Without algorithms, can you receive content you’re interested in, or represent your identity, values, or purpose? Without algorithms, do your posts reach your intended community? Without algorithms, do you have a digital community, or do you just have the void?

These networks have morphed into digital theaters. We come here to consume some form of entertainment. We’re routed to the appropriate theater and the appropriate seat based on our previous attendance, the duration we stayed, and how much money they made off of us. The difference is, the show never ends. Low friction, compressed choices, removed needs to exert effort, and optimized to keep us there.

Reclaim your lens

I’m writing this to advocate for a different perspective as we continue to build social networks.

Keep the platforms, but reclaim your lens.

A lens is not a place; it’s a way of seeing. It’s personal, transparent about its distortions, and easy to swap for another. With more lenses and the proliferation of social protocols, most of the problems we experience and the complaints we have about social networks will disappear.

A brief history of walls

Early applications (AOL, CompuServe, etc.) were walled gardens: curated spaces where everything “just worked” for many people. They took very new and intimidating technology and abstracted it for mass adoption. Tim Berners-Lee presented something more modest but yet more powerful: Hypertext Transfer Protocol. We’ve needed protocols since we began exploring networking and internet solutions (ARPANET, InterNIC, IANA, IETF, etc.), and as computer adoption increased over time, we needed better communication and data transfer protocols that could point to any individual's address no matter where they were in the world. Protocols like HTTP solved this; they were thin, public agreements that let anyone address, publish, fetch, and send information.

Protocols opened the world to new forms of discovery (web browsers, search engines) but we lost our sense of curation (Yahoo, etc.). Enter a new form of walls. Social networks (Friendster, Facebook, MySpace) solved curation, reputation, and speed at scale while enabling discovery – all in one place. They eventually collapsed our choices into a single canonical feed. Over time, the feed became the world. It’s efficient. It’s addictive. And it quietly trains us to outsource judgment.

Quiet constitutions

Protocols don’t tell you what matters; they just keep things moving and reachable. SMTP moves mail. HTTP moves documents. AT Protocol and Nostr move identities, content, and audiences. These are constitutions, not governments. They protect the possibility of many publics instead of asserting one public.

When importance is decided before movement and reachability – essentially, when data ranking importance precedes data access importance – you get speed and entertainment, but at the expense of permissionless systems coupled with a single, brittle sense of “what everyone thinks.” Protocols restore the order of importance: first data accessibility, then data rank.

Attention, scarce by design

Our attention is limited, compounding when invested well, and wasted easily. Social networks are excellent attention allocators. They optimize for engagement because that’s the currency they use to sell advertisements. The result is a culture of urgency, fueled by opaque rankings.

Ranking is power. It is tempting to label that power as “neutral math” (“the system simply surfaced this piece of content”), but ranking is a series of values and decisions: recency vs depth, breadth vs niche, novelty vs reliability. We have no “view from nowhere”. We have no “view without ranking”. We have no true neutrality.

We don’t need to reject the algorithms entirely; algorithms are extremely useful and can provide a positive experience. We need sovereign lenses: tools that let each of us decide how to spend attention, with clear trade-offs and visible costs.

Not what, but how

A lens is a commitment to how you see, not what you see.

• It’s personal. Tuned to your tolerance for novelty, dissent, depth, or speed.
• It’s legible. You can read your settings like a prescription; recency, diversity, domain balance, and topic focus.
• It’s swappable. Different lenses for work, for learning, for curiosity; potentially different lenses from different people you trust.
• It’s honest. It says, “I amplify recency; I down-weight links and replies,” instead of pretending to be neutral.

A good lens returns you to the web as it is: plural, uneven, alive. Like protocols, lenses aren’t places, they simply point to places.

What decentralization should mean

Simply put, “not centralized”. Controlled by many entities, not just one. This does not require blockchain.

• Multiple doors, one key ring. You can produce conversations anywhere without asking for permission.
• Many lenses, no mandatory doctrine. No single feed pretends to be reality.
• Context on arrival. Posts are shown with their place of origin so you see the house you’re entering.
• Leaving is easy. The primary action is “go to the source.” You go to the conversation; it isn’t extracted from its home.

Not abolishment

You don’t have to return to raw directories and bookmarks to reclaim the web. The point is not to abolish the feed; it’s to pluralize it. Feeds are lenses. A lens can be owned.

The open web already contains everything a “platform” promises: conversation, community, continuity. What it lacks is interoperability and habit forming rituals: treat the interface as a lens on top of places, not a place that replaces places.

Identity homes

Portable identities and social graphs (DIDs, etc) push decentralization down to the person. That’s helpful, but the core discipline holds even without perfect portability: let speech live where it’s spoken, and let identity be referenced, not reissued.

What we should resist & be cautious of

• The lens of truth. If everyone sees a single lens, it’s dogmatic.
• Concentration of truth. If most items only come from a few giants, lenses will essentially amplify narratives optimized for engagement, not aggregate collective consciousness.
• Opaque curation. If we can’t see or change the weights, lenses are asking for trust they haven’t earned.
• Extraction without return. If communities don’t get traffic back from the lens – or recourse when misrepresented – they will close their gates to the lens, impacting reachability.

Simple checks for lens builders

• Continuity. If your interface disappears, do the conversations remain accessible?
• Consent. Can a source opt out, or tune how it’s presented?
• Explainable. Can a reader look at an item and know why it’s there?

If you can’t answer “yes” to these, you don’t have a lens.

What the experience could feel like

You open a lens labeled “Deep Work.” It favors long threads with many distinct voices, slow half-life, high cross-source variety. You read a summary of a topic, then click through to the origin. Later, you switch to your “Fast” lens; short half-life, stricter diversity caps, stronger novelty bonus. Before bed, you try a friend’s lens called “Against My Priors.” It’s uncomfortable – in a useful way.

• You wander again, but with better guidance. Trails are optional, not required.
• You learn to compare lenses (recency, depth, dissent, etc.) rather than obey the one “true” feed.
• You see the place the content stems from: the forum, the board, the domain, etc. Many places, owned by many people.
• You leave more often.
• Your sense of “what everyone thinks” softens. There is no global agora, only overlapping publics. Answering these questions must be met with legitimate, quantitative attempts.

Same web. Different prescriptions. Different sources. Your choice.

A guide for developers

• Point, don’t possess. Always attribute, always link back.
• Choice over dogma. Lenses are easy to adjust, save, share, and compare.
• Prefer agreements to permissions. Wherever possible, follow open protocols.
• Publish your ethics. Not just a policy, a page that explains your trade-offs.
• Welcome competition.

The web you can still have

The open web never promised comfort, but it provided possibility. Walled gardens gave us comfort at the expense of sovereignty. The antidote is agency: keep the power of algorithms while returning judgment to the user, keep the power of aggregation while returning ownership to the author, and keep the power of competing enterprises while remaining decentralized.