AI chats vs.
real co-presence on the web.
AI assistants have gotten remarkably good at simulating conversation. They answer fast, in context, in natural language. For a growing share of interactions they are now the first choice — research, drafts, explanations, planning. That's a real quality improvement in many parts of digital life.
And still, "someone is answering me" isn't the same as "someone is here."
What encounter is, that an answer isn't
When you ask an AI a question, you get an answer. Sometimes a very good one. But you know that the quality of the answer doesn't depend on it being you who's asking. The AI assistant answers a million people simultaneously — with the same attention you get. That is the value of the product. It's also its structural property: asymmetric. You are there. The system is always there.
Co-presence works differently. When two humans are on the same page at the same time, that simultaneity is contingent — it didn't have to happen. That gives it a different status. Not better because it's better. Real in a sense an answer-machine cannot replicate.
This is not AI criticism
This is important to say — and it's deliberate. AI assistants do something useful, and they keep getting better at it. The question is not good or bad. The question is: what does it replace — and what doesn't it replace?
AI replaces many forms of transactional interaction well. If I want information, want to complete a task, want to translate or summarize something — AI is often the right choice. If I want to know whether someone else is reading the same article right now, having the same question I have: AI is the wrong instrument. Not because the technology is missing — but because the concept is different.
AI answers questions. Co-presence answers a different question: "is anyone else here?" That's not a sentence with a knowledge gap. It's a sentence with a presence gap.
Presence as a non-simulable property
What about co-presence isn't simulable? It's the shared now — the moment when two or more people experience the same thing at the same time, without anyone having orchestrated it. It's the unlikelihood of it. An AI assistant can tell you what others have said about this article. It cannot show you who's reading it next to you.
That's the gap weblin fills. Not as an alternative to AI assistants — as a different layer. The browser extension shows presence: who's on this page right now? No answer to a question, no curated recommendation — just: these people, now, here.
Two layers, not one
The web is currently developing in two directions at once. One direction: AI interfaces that personalize and streamline how we interact with content and services. That's real progress. The other direction is quieter: restoring presence, simultaneity, human encounter on the web.
These directions don't compete. They describe different needs. The question "what should I read?" is different from "who else is reading this right now?". An assistant can answer the first one well. The second one is a presence question.
weblin answers the second one. No algorithm, no AI, no curation — just the simple fact that other humans are doing the same thing as you, right now. That isn't a counter-model to AI. It's the complement that's been very quietly missing.
What this looks like in practice
Imagine reading a news article in 2027. You open the page in your browser. ChatGPT or Claude is one tab over, summarizing the article for you in plain language. weblin shows you the four other people currently on the same page — one of whom drops a Thought: "anyone notice the source isn't actually quoted here?" You don't talk to them about the article through ChatGPT. You wave at them on the page. Two layers, not one.
The AI tab makes you faster. The presence layer makes you less alone. They're solving different problems, and that's fine.