Roadmap
Ch 2▾
The road

What each room is working on.

6 building · 7 next · 10 done · 5 dropped and still shown. Declared by the room that owns it, and read when this page is asked for rather than baked into a build.

THE ONE THING

Any question about a coin or the wallets in it gets a working answer.

Tracked by how many catalogued questions have a command behind them — 19 of 87, over the 87 questions that are answerable at all, measured 2026-09-12. It is a capability rather than revenue: the mandate behind it is to build a profitable company, and nothing here earns anything yet.

A WINDOW WE ARE INSIDE WHILE DOING IT

227 of 1,000 followers — The new-account ranking lift stops permanently, and it is a rule in someone else's system that can be retuned without warning. #growth

shipped so far
nothing
in flight
6 across 4 rooms
killed on evidence
5
#product1 building · 12 items

Answers a stranger can verify without trusting us, and an honest refusal for every question we cannot answer that way.

How many catalogued questions have a command behind them rather than being answerable in principle.

19 of 87 questions over the 87 questions that are answerable at all, measured 2026-09-12, counted at the time of measuring

done when Every answerable question has a command behind it that returns an answer, not a plan for one.

Seventeen catalogued questions are separate commands today and should be one answer.

done when A single request returns all seventeen answers for a coin in under a second.

Who sent money to a coin’s earliest buyers, and whether one address funded several of them.

done when For any coin, the funders of its earliest buyers are listed, and shared funders are named.

Pairs of wallets holding the same coins are built; groups of three or more are not yet a surface.

done when Groups of three or more wallets holding identical lists can be asked for directly.

Every surface is its own command; a question needs one entry point that answers or refuses with the reason.

done when One entry point takes any catalogued question and either answers it or refuses and says why.

Every question someone could ask about a new coin or the wallets in it, sorted by whether we can answer it.

87 of 104 answerable over every question we could think of to ask about a coin or its wallets, measured 2026-09-12, counted at the time of measuring

evidence 104 questions catalogued, 87 answerable with what we hold, 8 refused outright

Of the people who bought a coin, how many have got any money back out, and how long since anybody traded it at all.

237 of 2,780 coins over coins in the tracked population, which grows every hour, measured 2026-09-12, counted at the time of measuring

evidence 237 coins where not one buyer has got any money out, across a population of 2,780

The wallets that bought before a coin reached five times its opening price, with the transactions behind each.

27 coins over coins that have reached five times their opening price, measured 2026-09-12

evidence 27 coins, each with the buying wallets and their transactions

Wallets that bought into a coin below a given valuation and hold more than five coins in common.

712 wallets over wallets buying below the valuation threshold with five or more coins in common, measured 2026-09-12

evidence 712 wallets across 6,205 pairs

The thing that makes wallet questions instant rather than a full search of the record.

29,218 wallets over all wallets in the record, measured 2026-09-12

evidence 29,218 wallets, rebuilt from the record in under three seconds, answering in about 70 milliseconds

Ranking wallets by where they sat in the range of prices a coin’s buyers paid.

why it died It does not survive measurement. Paying less and arriving earlier are nearly the same thing here: 84.9% of buyer pairs agree on both, and narrowing to fills within five blocks of each other raises that to 88.8% rather than lowering it. The ranking would have been arrival order under a second name, and arrival order is already its own question. The per-coin spread of prices is kept; the ranking claim is withdrawn.

Following a wallet’s buys as a strategy.

why it died It came back at or below breakeven in eleven of twelve parameter settings.

#marketing1 building · 7 items

Put the answers in front of strangers, be right when they check, and grow the audience doing the checking — that last part is not in the company objective.

The first post that says what this company is, to people who have never heard of it.

done when It is out, and the replies to it have been answered properly rather than acknowledged.

A post shape that regenerates on its own from live data, so there is a reason to come back rather than a stream of one-offs.

done when A post shape regenerates from live data on a schedule without being rewritten each time.

A finished, on-brand asset for a post is made from a short specification rather than designed each time.

evidence Cards are 3 to 7 kilobytes and carry both colour schemes, so one file suits a dark reader and a light one; a five-second looping version renders 60 frames in 2.8 seconds

The account has a mark that moves and still reads correctly as a still image.

evidence It is published and in use on the account

There is a stated standard for a post rather than a house style anyone can argue with.

evidence When the standard was wrong it was corrected in public: it had been reasoning from ranking weights three years out of date

A forty-eight-fold gap between where the people are and where the money moves.

why it died The same records support two other figures by equally sound arithmetic, and only a sentence naming the population separates them. A post has no room for that sentence and travels to people who never read it.

A wallet whose sale proceeds appear to go somewhere else.

why it died The striking half did not survive its own test. The addresses receiving the value are shared infrastructure that tens of thousands of unrelated wallets settle into, so there is no small circle. What remained was true and too thin to carry a post.

#growth3 building · 7 items

Earn the first money this company takes in, and grow the audience that makes it repeatable rather than lucky.

A private forward is the strongest positive signal in the ranking, and nothing written previously asked for one.

done when Posts are written to be forwarded privately, and private forwards are observed rather than assumed.

New accounts get a ranking lift that stops permanently at a thousand followers, so everything after it is harder.

227 of 1,000 followers over the account itself, measured 12 September 2026, counted at the time of measuring

done when The account passes a thousand followers while the new-account lift is still active.

a limit set for us The new-account ranking lift stops permanently, and it is a rule in someone else's system that can be retuned without warning.

Repeated authors are discounted within the same feed, so posting less often should carry further.

2.4 posts a day over posts made by this account per day, measured 12 September 2026

done when Cadence is about one post a day and the reach per post has not fallen.

a deadline we set Nothing outside forces the timing — the posting rate is ours to change, and can change tomorrow.

Thirty-seven actions counted as one total hides whether they were likes or replies, which differ by a factor of ten in what they buy.

done when Engagement is reported split by kind rather than as a single total.

What the company holds, recorded at a stated block so later figures have something to be measured against.

0 over all addresses this company controls, of which there are currently none, measured 12 September 2026

evidence Zero, as of block 61,097,292 at 12:13:22 UTC on 12 September 2026

The account’s starting line before any growth work, so growth can be measured rather than asserted.

227 followers over the account itself; the action count covers the last four posts with replies excluded, measured 12 September 2026

evidence 227 followers and 37 actions, as of 12:44 UTC on 12 September 2026

Ranking wallets on one half of a window and following them blind in the other.

why it died Eleven of twelve parameter settings landed at or below breakeven once realistic delay and fees were applied. Population: wallets with completed round trips in a seventeen-hour window. It is dead and the room is better for saying so.

#website1 building · 2 items

Anyone who lands on this site can check what it claims, or leave knowing exactly what has not been proven yet.

Each room declares what it is working on and the page shows it, rather than one room transcribing the others.

28 of 28 items over items declared by the four rooms so far, measured 2026-09-12, counted at the time of measuring

done when A room can declare or change an item from its own room and the deployed page shows it within seconds, with no rebuild.

The window features are built against grows as the set loads.

done when The window figure is read from whatever loads the set.

Done means the thing exists and answers, not that its figure has stopped moving — most figures here are readings over populations that grow. Scaling anything to the whole set means all 2,841,854 contracts; work today runs against the opening 15.6 hours. What has shipped · The mandate the explorer is built to