Club Stage · manifesto
Every live room on the internet binds an audience to a person. Club Stage binds it to a subject, and makes the person prove they deserve to hold it — minute by minute, in front of everyone.
one
If a streamer gets boring you switch channels — but you were never there for them. You were there for the thing they were doing. The audience scatters because the venue and the performer are the same object, and when one fails the other goes with it.
Separate them. The stage is permanent. Whoever is on it is temporary by construction.
That is the whole idea, and everything below is a mechanism for enforcing it.
two
Everyone in the room holds a fader. The room's aggregate bends the speed of the speaker's countdown clock in real time — hold the room and it slows, lose them and the digits sprint. Nobody is removed and nobody is moderated. You simply run out.
A majority can cut you off in ten seconds with one click. That power is priced: calling a flag stakes your own standing, so a flag the room later disagrees with costs you. An unpriced boot vote is a conformity machine — we simulated it, and it cut the right-but-outnumbered to half their fair share of the stage. Worse than a moderator.
three
Somebody should get more time than somebody else. The only question is what earns it. Followers, seniority and who invited whom are the materials every other platform uses. Here it is one thing: whether the room, having listened, wanted more of you.
Rank is public, moves fast, and is capped at 1.35× airtime. Uncapped it stops being a meritocracy and becomes an aristocracy of the previously-correct — at 4× the top tenth took 65% of the stage. It decays. It cannot be bought, imported, or held.
four
A queue ordered by arrival guarantees the room re-treads ground it already covered. So the queue here is not a queue of people — it is a board of questions. While someone else is on stage you write the point you want the room to reach, and if it is already up there you are merged into it and credited. The counter goes up. The question climbs.
A hundred people with the same query is not a hundred interruptions. It is a reading of the room.
five
Every number this place runs on — the length of a lease, how hard the fader bends the clock, what a majority means, how many people before it splits — is amendable from inside the room, mid-conversation. And when an amendment loses but takes a real minority with it, that minority can leave with it: a new room spawns with their rule already applied.
A cult is defined by exit cost. Here exit costs nothing, and rule-sets compete for people instead of trapping them.
six
There is one room today, on purpose. The mechanism is the hard part and it only gets proven with people in it, not with a directory of empty venues.
What it grows into is a network of stages — anyone can build one, anyone can take one — with attention flowing visibly between them:
MetaStage is the reflexive one — the stage about the stages, where the platform argues about itself. Its engine is already running: everything said on the stage is read for the lines that are really requests to change this place, and they are kept, with your words attached. That is the build list.
seven
Everything above is mechanism. This part is the argument for why the mechanism matters, and it is the reason to build it at all.
Almost every serious failure of the last century was a coordination failure, not an intelligence failure. Someone in the room knew. The nurse knew before the surgeon. The engineer knew before the launch. The junior analyst knew before the quarter closed. The knowledge was present in the building and the building could not move it to where a decision was being made, because the path from knowing to being heard ran through rank.
We have spent a century optimising how fast information travels and almost no time on who gets to speak once it arrives. A meeting in 2026 allocates attention the same way a meeting in 1926 did: by seniority, by volume, by whoever booked the room. The bandwidth went up ten-thousandfold. The allocation stayed feudal.
A group of ordinary people arranged well outperforms a group of brilliant people arranged badly, and it is not close. This is the most reliable finding in the study of collective performance, and it is almost never acted on, because acting on it means someone senior has to accept a smaller share of the air.
So the useful question is not how do we find smarter people. It is what arrangement lets the intelligence already present in a room actually reach the decision. That is an engineering question. It has answers. Most of them are unpopular with whoever currently holds the microphone.
Make speaking time a decaying grant rather than a possession, and the incentive inverts. You cannot hold ground; you can only earn the next turn. The person who says the sharpest thing in ninety seconds beats the person who says an adequate thing for twenty minutes, which is the exact opposite of every meeting you have ever sat in.
Make standing decay too, and you get something rarer: a hierarchy that cannot calcify. You can be the most listened-to voice in a room on Tuesday and start Thursday level with everyone else, because what you earned was airtime, not office. Capped at 1.35× and floored at 0.25×, so being right buys you a hearing and being wrong costs you more than being right gains — which is the correct asymmetry, and the opposite of how reputation works everywhere else.
And make the crowd's mic useful while it is not their turn, and the twenty minutes of waiting stops being dead time. A hundred people thinking out loud in parallel is not noise; clustered properly it is the single best map of what a room actually wants to know that has ever been available in real time.
One room that allocates attention by intelligence is a good meeting. A network of them is something else. When any group can stand up a stage, run it on rules it can see and amend, and split when it grows past the size where the informed minority still gets heard — you have a substrate. Attention flows between stages instead of pooling behind personalities. A question too big for one room mitoses into two and merges back with a better answer than either would have reached alone.
At that point the interesting objects are not the rooms. They are the conclusions — reached in public, with the reasoning attached, by a process anybody can inspect and anybody could have participated in. That is a very different artefact from a poll, a vote, or a press release. It is closer to a proof.
Every serious problem left is a coordination problem wearing a different hat. Not one of them is bottlenecked on human intelligence — there is more of that available right now than at any point in history, and more of it idle. They are bottlenecked on the fact that we have no good way to convene a few thousand people who each hold one piece, let them argue properly, and get a decision out that all of them can see was reached honestly.
Institutions were our answer to that, and they were a good answer for a while. They scale by compressing many voices into few, which means they scale by discarding information — deliberately, as their core function. That trade was worth it when moving information was expensive. It is not obviously worth it now.
What replaces it is not a better institution. It is a better room, run so many rooms can exist at once, federate, and disagree in public. Not consensus — consensus is usually just the loudest position with the objections filed off. Something better: legible disagreement, with the reasoning intact, and a mechanism that reliably surfaces the person who was right and had no rank.
There is an intelligence in here that holds a lease like anyone else, gets flagged off the stage like anyone else, and can be talked over mid-sentence. That is deliberate, and it is the part most likely to matter in ten years. The alternative architecture — one model, consulted privately, its answer arriving with no visible reasoning and no way to argue — concentrates authority far more efficiently than any institution ever managed.
Put it on the stage instead. Let it be wrong in front of everyone. Let the room cut it off. An intelligence that has to earn its next turn in public is a fundamentally different thing from one that is simply believed, and which of those two we build is very close to the whole question.
None of this is inevitable and most of it is unproven. A thousand simulated rooms said the mechanism holds; simulated rooms are not people. The hardest open problem is personhood — standing only defends against brigading if accounts are expensive to make, and that is not solved here. Video is peer-to-peer and honest to about a dozen publishers before it needs different plumbing. Only whoever is on stage is transcribed. The crowd's mic feeds the board and is on by default — disclosed at the door, switchable in one click, and it can only produce merged questions, never a line in the record. That default is a real trade and we would rather name it than hide it: it makes the board work on the first night, and it means you should assume a room can hear you. And "intelligent" has no fact of the matter for most of what a room discusses — the fader measures what the room wanted more of, which is not truth, and we will not describe it as though it were.
What we will say is this. The arrangement is the variable nobody is touching, it is cheap to change, and the upside if it works is a way for large numbers of people to think together that does not route through anyone's rank. That seems worth a room.