Why the web
needs rooms again.

There's a version of the early internet that everyone remembers slightly differently — but in nearly every memory, there are places. Forums that felt like local pubs. IRC channels where you knew who was around. Webpages with personality, because someone had built them and was usually living inside them too. The web had rooms. Imperfect ones, but recognizably spatial: you entered something. You met someone. You could leave.

That was a while ago.

What timelines did to places

Sometime between 2008 and 2012, platforms reorganized the architecture of the web. Not through any single decision, but through a slowly dominating pattern: the timeline. A stream of content, sorted by recency or — later — by relevance score. The benefit was real: more content, faster to consume, always current. The cost was structural: rooms got replaced by streams.

What was lost is hard to name, because it isn't measurable. It's not a feature you'd miss in a product changelog. It's more like this: when you open a webpage, you're alone. Always. You see what others have left behind — comments, click counts, likes — but you don't see who's currently here. Presence has disappeared from the web.

There used to be cues. The ICQ flower turned green when a friend came online. IRC's /who command listed everyone in the channel. A vBulletin forum showed "12 users currently viewing this thread." Even MySpace had the awkward "Top 8" — a way of declaring co-residence. Then, one product cycle at a time, those signals were stripped out. Nobody mourned them, because nobody named what was being lost.

What makes a place a place

In the 1980s, the urban sociologist Ray Oldenburg described what he called third places — locations beyond home and work that serve as social anchors. The café, the pub, the bookshop, the barber. What defines them isn't primarily their function. It's their simultaneity: multiple people, at the same time, in the same space. Not synchronized by a schedule. Not curated by an algorithm. Simply there.

Three characteristics matter. Accessibility: anyone can walk in, no formal membership required. Neutrality: no one has home advantage; the space belongs to none of its visitors individually. Regulars: people return — not because they have to, but because the place exists for them. These three together produce something that doesn't exist without them: an atmosphere, a sense that you belong to something simply by being there.

Most digital products have translated this concept into messaging features: chats, groups, channels. That's closer — but it stays a separate layer. You don't visit a webpage and find yourself with others on it. You leave the webpage and open Discord.

The defining property of a third place isn't what you do there. It's that other people are there too, doing nothing in particular, in the same space, at the same time.

Why feeds are not rooms

A feed is not a room. It is a stream. The metaphors give it away. You don't enter a feed; you scroll through one. You aren't in a feed with anyone; you're moving past what they posted. The grammar of a feed is the grammar of consumption. The grammar of a room is the grammar of being.

This isn't an aesthetic preference. It's an architectural difference with behavioral consequences. A room produces encounter; a feed produces exposure. A room invites returning; a feed rewards being there now. A room can have an atmosphere; a feed has volume. None of these are bad in themselves — they just produce different things. The mistake is thinking we can have one and not need the other.

What the feed-architecture optimized for

Feeds didn't replace rooms by accident. They replaced rooms because they scale. A room has natural capacity limits — atmosphere collapses past a certain density. A feed scales linearly: more content, more users, more ad inventory. For a business model funded by attention, feeds are the obviously superior architecture. For a model of encounter, they are roughly the worst.

Presence is not nostalgia

It would be easy to read this as a retro project — a longing for the early web that, in fact, is a longing for one's own youth. That would be a wrong reading. Presence is not retro.

Presence is why people book coworking spaces despite working remotely being available. Why libraries are fuller than they've been in decades, even though every book is digital. Why concerts still happen, even though Spotify costs €11 a month. The simultaneous being-there with others has a value that isn't replaced by better content. It's a different kind of value.

The pattern is consistent: the more efficient remote alternatives become, the more people pay for the inefficient simultaneous ones. That isn't because remote is broken. It's because presence and content are not the same axis.

Why digital tries failed (so far)

There have been attempts. Virtual worlds: Second Life, Habbo Hotel, later VRChat. They got something real — the longing for presence — but they packaged it into a separate environment you had to enter. You weren't on the web with others; you left the web for another world.

Discord is closer: servers, channels, that feel like rooms. But Discord is still a separate platform. The café isn't where you're working — you have to go to it. Slack, Teams, Mattermost share the same shape. They're rooms, but rooms in their own building. The web stays unpopulated.

The earlier attempt — the original weblin in 2007 — got the architecture right but the timing wrong. Browsers were slow. Privacy primitives were primitive. Mobile barely existed. The 100,000 users we had at peak weren't enough to overcome the technical drag. We sold to Sulake in 2008, the project quietly dissolved, and the question dissolved with it.

What weblin does differently

weblin asks a different question. What if the room weren't separate? What if every URL — a news article, a GitHub repository, a product page, a forum thread — could also be a place where people are simultaneously present?

This isn't metaphor. weblin is a browser extension — Chrome, Edge, Firefox, with Safari and mobile coming with the Kickstarter funding — that sits as a layer on top of every webpage. It doesn't rebuild the page. It doesn't require integration. It doesn't replace anything. It just adds one piece of information that platforms removed: who else is currently here.

Anyone visiting the same URL at the same time appears as an avatar along the bottom of the page. You can wave. You can send a short message anyone on the page can see (Thoughts, which fade after 7 days — no permanent harassment archive). You can ignore someone on this page, or block them across all pages. URLs are hashed on your device before they leave; we literally can't see what you're reading. The mechanism is open source, on GitHub, audit-able.

The room is the webpage you were already on. There's no platform- switch, no new identity to maintain, no new feed to check. The presence layer is co-located with the content layer. That's the move.

What changes when rooms come back

The product is small. The implications are not. A web with presence is a web where:

• Reading the same article as someone else stops being a coincidence and starts being an encounter.
• Indie websites can host community without building their own Discord — the community is on the page.
• Festivals, magazines, niche fandoms get co-presence on their own URLs, not on someone else's platform.
• The web stops being a library full of empty reading rooms and becomes, again, a library where you sometimes look up and see who else is here.

The web was never designed as a feed. It was designed as a collection of places. Maybe the right question isn't how to navigate content faster — but how to arrive somewhere, again.

One layer, no rebuild

The reason this is possible now, when it wasn't in 2007, is that nothing has to be rebuilt. Browsers are fast. WebExtensions are standard. Real-time presence over XMPP-style protocols is well-understood. Privacy primitives like client-side hashing exist and work. Open source infrastructure is normal, not radical. Platform fatigue is real, not theoretical. And awareness of platform-dependency has been hammered into the public by a decade of API shutdowns and algorithm rugpulls.

The idea is the same. The conditions are different. The web is still here, mostly intact, mostly underpopulated. weblin doesn't replace it. It just turns the lights on and shows you who else is in the room.