A founding team that actually argues about your idea

How a chapter runs

A venture is made of chapters, and a chapter is one bounded debate: its own rounds, its own wrap-up, its own score. The first one is the opening assessment of the idea — this is how it goes.

1. Bring an idea — even a rough one A sentence or two is enough to start. The team asks a couple of quick clarifying questions first if the idea is too vague to argue about productively.
2. The Opening: independent first takes The five core specialists are at the table from the start — there's no line-up to pick. If the idea turns on something none of them cover, the Program Manager also seats a domain specialist before anyone speaks, and can propose bringing one in later if the room keeps hitting the same gap; that one's your call to accept or decline. You can also bring one in yourself at any pause — type /agents in the box where you'd normally comment. The Opening isn't a debate: each person writes their first read at the same time, without seeing the others, so you get real opinions instead of everyone agreeing with whoever spoke first. That's why it's called the Opening and the debate rounds count from 1.
3. Then they actually argue, one at a time From Round 1 on, nobody writes in parallel. Each person speaks in turn and reads everything said before them in that same round — so a question put to the Financial Expert by name gets answered there and then, not left hanging. That's the difference between a debate and five essays stapled together.
4. The Program Manager decides who's needed It reacts to exactly what was said — naming the real disagreement, attacking comfortable consensus, calling out vagueness — then puts a specific question to the two or three people who can actually move things forward, and leaves the rest out of that round. The relevant voices engage; nobody pads the room with a turn they had nothing to add to.
5. You can cut in mid-round You don't have to wait for a round to finish. Send a comment while the room is talking and anyone who hasn't spoken yet reads it and responds to it directly — it takes priority over whatever they were asked. Between rounds you decide whether to run the next one or wrap up; nothing advances without you.
6. Wrap up, and get a real verdict A summary of where the room landed, concrete next steps to actually go do, and an Idea Scorecard — not just a transcript. That closes the chapter. You keep the room: you can start the next chapter whenever you have something new to bring it, or close the venture and stop there.

Then you go and find something out

A debate can only work with what you bring it. The most useful thing you can do after a chapter is leave, test the thing everyone was guessing about, and come back with an answer. That's what the next chapter is for — and the room remembers: every decision the venture has settled, the previous chapter's wrap-up and score, and the question you opened this one with.

You pick what kind of chapter it is, which changes how the room argues — each kind carries its own instructions to the room, handed to every seat when the chapter opens. The kinds that ship, which an administrator can reword or add to:

Every chapter stays readable. Go back to an earlier one at any time to re-read the argument, its decisions and the score it reached — it opens read-only, so there's nothing to break by looking. When you're done with the venture, close it and say how it went: you're pursuing it, you decided against it, or you're parking it for now. Parked isn't deleted; reopen it whenever it comes back.

Who's at the table

Those six are in every room, and they are the floor rather than the ceiling.

Taking one of them aside

Some things you can't say in front of the whole table. The real runway rather than the one in the deck. A co-founder you're worried about. Someone who might leave. You can open a private conversation with any one specialist, and it runs alongside the debate rather than pausing it — the room carries on without them while you talk.

What you say there reaches that specialist and nobody else. It doesn't go into the transcript, the scorecard or a report. What it does is change how they argue from that point on: a Financial Expert who knows your real runway pushes for a shorter path to revenue without ever saying why, and the others answer the argument on its merits, because that's all they can see.

That matters for what the scorecard means. The room is still the whole record — a private briefing only reaches the outcome through arguments the room can read and push back on, the same way a specialist's own expertise does. Nothing is scored that wasn't argued.

Only you can see any of this. A venture you've shared read-only shows no sign that a private conversation exists.

Domain experts, when the idea needs one

The six cover the shape every venture has: a market, a product, a way of reaching people, a set of numbers, and something to build. What they don't cover is your industry. A food truck turns on kitchen throughput and licensing. A clinic turns on data residency and who is allowed to sign off on a diagnosis. A freight idea turns on what a customs broker actually does. None of that is general business knowledge, and a room that pretends otherwise gives you confident answers about things nobody at the table has ever done.

So the roster extends. A domain expert is a specialist in one particular ground, who joins mid-debate and argues alongside the six for the rest of the venture. They read the whole discussion before speaking, the Program Manager briefs them out loud on where it actually stands, and they answer the question that justified bringing them in straight away — an expert who joins and says nothing until the next round is just a more expensive seat.

Two ways one arrives

The room asks for one. When a question the discussion keeps returning to sits outside everyone's expertise — usually because somebody in the room has said as much themselves — the Program Manager proposes a specific specialist, names the moment that exposed the gap, and tells you what it would put to them the second they sit down. You decide. It proposes nobody on most rounds, and never to cover ground the room is already handling well.

You bring one in. Type /agents while the room is paused and choose from the roster yourself. There's no approval step, because you're the one asking. Use it when you know something about your own industry that the room hasn't worked out yet.

Why the room doesn't just seat everyone

An unnecessary specialist makes the room slower and each round more expensive, and a debate with twelve voices is worse than one with four — people talk past each other rather than to each other. Most ideas need no domain expert at all, so the absence of one is a judgement the facilitator made, not an omission. Every seat is billed for each round it speaks in, which is why a room has a size limit rather than growing indefinitely.

The roster of available experts is not fixed either. It grows from what rooms actually turn out to need: when several different ventures keep leaving the same kind of question unresolved, that gap is a candidate for a new specialist — reviewed by a human before anyone can be seated in a real debate.

The Idea Scorecard

Every time you wrap up a chapter, an evaluation pass — separate from the debate itself — scores the idea honestly across six fixed dimensions, each scored 0-100 and cited against something specific that actually came up. A seventh, Domain Depth, appears when a domain specialist sat in.

Each chapter is scored on its own evidence, so the scores are comparable to each other and you can watch a number move as you learn things. None of them is a running total, and a later chapter doesn't overwrite an earlier one's verdict.

Scored the way a fund would

Every fund scores ideas on its own matrix. So a venture can be scored on an evaluation framework: the house scorecard above — the default, and what you get if you choose nothing — or a framework modelled on how funds at a stage tend to score. Three come built in, each labelled “in the style of” because they are modelled on public writing, not on any fund’s actual matrix:

A framework has its own dimensions, says what counts as evidence at its stage, carries the fund’s weights, and may have gates: a dimension a fund will not average away, where a score under the line makes the verdict “not yet” whatever the rest scored. The room is told what the venture will be scored on, so the specialists argue those dimensions; the evaluator scores exactly them. A dimension the room never examined is shown as not examined, never as zero.

You choose the framework once, when the venture is created, and every chapter is scored on it — that is what keeps a score comparable from one chapter to the next. The room header shows Scored on: …. Any finished chapter can be re-scored under another framework from the bar beneath its scorecard: the same debate read again, no new rounds, one small charge, the chapter’s own score untouched — and its PDF downloaded under that reading. That is how you read a chapter the way a particular fund would before you pitch them.

Two reports

Both are formatted PDFs, built the moment you ask for them rather than filed away at wrap-up — so a download is never a stale copy of something the room has since moved on from.

What this isn't

It's a decision-support tool for thinking through a startup idea faster and more rigorously than arguing with yourself — not financial, legal, or professional advice, and not a guarantee that a high score means the idea will succeed or a low score means it won't. Treat the debate as a sharper starting point, not a verdict to outsource judgment to.