A Miranda rule for founders — and for futurists-in-residence, chief believers, heads of curiosity; not for highly effective managers and other operators
The Miranda warning is filed under criminal procedure. It belongs in a grammar book.
"Anything you say can and will be used against you" is not just a rule of evidence. It is a rule of conjugation — the only one ever written down for a verbal mood that officially does not exist, although every founder speaks it fluently and most investors are paid to pretend they can hear the difference. English textbooks list the future indicative: the mood in which we state facts about what will happen. Delete a single letter, and the mood that states facts becomes the mood that enters them into evidence: the future indictive. The indicative states facts; the indictive states facts that can and will be used against you.
Other names auditioned for the role. The future accusative — the Latin tailcoat; it carries an etymological bomb we will detonate below. The indictable future. The prosecutorial future perfect — "he will have lied." The future anterior of fraud, for readers who take their venture capital with continental philosophy. But the honest name is the minimal pair, because the whole phenomenon is a minimal pair: one letter between a forecast and an exhibit.
Everything you say in the future indicative may one day be read back to you — in the future indictive.
Not the past, not the present, not quite the future — a mood for which ordinary language has no honest word. A startup is not a lie about the present. It is a wager that the future will arrive before the present demands proof.
The founder says: we are building a bank. There is no bank. He says: we will get the licence. No one has issued it. He says: this will be an enormous market. The market, at present, consists of twenty slides, three meetings, two sceptical partners, and one person who has inexplicably not yet given up. If it works, those sentences will later be called vision. If it doesn't, the same sentences will be called misrepresentation. The future, like a good lawyer, arrives only for the winners.
We are bad at telling prophecy from pitch — a pitch is a map drawn before the territory — bad at telling pitch from simulacrum, simulacrum from fraud. So we wait for the final whistle, cheer retroactively for whichever team won, and call morality what was only the score on the board.
The industry admits all of this in its paperwork. The deck is not a report on the present; it is a stage set of the future. Its foundational instrument is literally named for the mood: a SAFE — Simple Agreement for Future Equity — a security whose entire content is a verb tense.TAM. Roadmap. We are building. We are in talks. We can become X. This is the vocabulary of probabilistic reality — and it is legal, and it has to be, because if you outlaw the grammar of dreaming, early venture does not get cleaner. It just stops. A purely present-tense startup is, in any case, unfundable — we currently have a bug, a prototype, a Stripe account, and an argument is not a seed round; it is a diary entry. The founder's job is to make the future gravitational before it is factual.
Now the etymological bomb. The accusative case — the case every schoolchild memorises — is a two-thousand-year-old translation seam.
The Greek grammarians called it ptōsis aitiatikē, from aitia, a word with two meanings: cause and accusation. Aristotle's four causes are aitiai; a courtroom charge is also an aitia. By the standard account, the Greeks meant the case of the thing caused — the object an action brings about. The Roman grammarians picked the other sense and rendered it accusativus: the case of the accused. The Slavic languages, translating in their turn, accidentally preserved the whole ambiguity: the Russian винительный падеж comes from вина, whose older meaning is not only "guilt" but "cause" also — a sense still alive in everyday speech (по вине дождя: "by fault of the rain," meaning simply because of it).
In other words, Indo-European morphology has been trying to tell us something for two millennia: causation and accusation are one word. A founder causes a future. But if the future fails to arrive, the causing is reread — retroactively, effortlessly, in the same grammatical case — as accusing. Aitia cuts both ways.
Which is why founder is such a dangerous noun. A founder is not merely someone who begins; a founder is someone who causes. He bends resources toward something that does not yet exist, hires people away from safer jobs, turns cash into a thesis, persuades customers to tolerate immaturity in exchange for proximity to tomorrow — he asks reality for credit. When reality is repaid, we call him a visionary. When reality defaults, we call him liable.
Bulgarian has an entire verbal paradigm — the renarrative mood — for "allegedly; so they say; I did not witness it myself." Turkish does the same work with the suffix -mış. Bulgarian newspapers are expected to report unverified claims in the renarrative: the morphology itself discloses the epistemic status of the sentence. In those terms, the whole venture mechanism compresses into one line: investors fund in the renarrative and sue in the indicative; due diligence is the ritual for converting -mış into the witnessed. The data room is a chapel where reported speech is changed into declarative fact. Traction is what we call the future tense once it starts leaving fingerprints in the present. (In my own book, every load-bearing claim wears a tag — fact, inference, or what-if: a homemade evidential mood, bolted onto a language that doesn't ship with one.)
And one more gift from the world's grammars. Chukchi and the Eskimo–Aleut languages mark motion along a path with its own case — in their grammars it is called the prosecutive case, from the same Latin root as prosecute: prosequi, to follow along, to pursue. A founder lives in the prosecutive literally. It is the same road, named for movement while he keeps moving — and renamed for pursuit the moment he stops.
He would later flee to Moscow. She was to become the youngest self-made… defendant. It is the house tense of biographers, prosecutors, and hindsight.
J.B.S. Haldane described the four stages through which a new idea passes on its way to acceptance (1963): (i) worthless nonsense; (ii) interesting, but perverse; (iii) true, but quite unimportant; (iv) I always said so. The uncomfortable — and strangely consoling — footnote came from Baruch Fischhoff, who in 1975 demonstrated hindsight bias experimentally: once people know an outcome, they cannot reconstruct their own prior uncertainty. Memory silently rewrites yesterday's "surely not" into "I knew it all along." Stage four, in other words, is not (only) villainy. They are not lying when they say they always knew. Their brains cooked their past for them. Which produces the symmetry this book exists to point at:
Two artificial tenses, aimed at each other across a term sheet.
— first ridiculed, then violently opposed, then accepted as self-evident. Schopenhauer never wrote it; it appears nowhere in his works. The Gandhi version — first they ignore you, then they laugh at you — is not Gandhi either; the earliest documented form belongs to Nicholas Klein, a union orator addressing clothing workers in 1918. Carl Sagan closed the entire genre with one observation: yes, they laughed at Columbus and the Wright brothers — "but they also laughed at Bozo the Clown."
Ridicule, as an instrument, cannot distinguish early from wrong. Only timestamps can. Which is why the seventeenth century — a serious century — built timestamp infrastructure. When Galileo observed something strange around Saturn in 1610, he sent Kepler an anagram: priority staked, discovery undisclosed. When Hooke wanted to claim his law of elasticity in 1676, he published ceiiinosssttuv, and unfolded it two years later into ut tensio, sic vis. The anagram was the DOI of the seventeenth century.
And then the system produced its most baroque result. Kepler decoded Galileo's Saturn anagram incorrectly — into a sentence about Mars having two moons. No one had seen any such thing; no telescope on Earth could have. Two hundred and sixty-seven years later, Asaph Hall pointed a better instrument at Mars and found exactly two. A bug in the timestamp system compiled into a prophecy. That is the future indictive at its outer, cosmological limit: even a 'misreading' can be vindicated, if you give the future enough time.
The law, meanwhile, guards the past and the present — and is honest enough to admit what it does with the future: it exempts it. The safe-harbor provision for forward-looking statements (US, 1995) is the literal legalisation of the future tense: say "we will," say "we plan," wrap it in cautionary language, and the sentence is untouchable. An indictment, moreover, needs damage — and damage is the great translator. While the rounds keep closing, there is no injury and no complainant: Wirecard lived inside that grammatical shelter through five years of published questions. But when the music stops, everyone in the room becomes a linguist. Verbs are inspected; modal auxiliaries are exhumed; would is weighed against will; in discussions is placed under fluorescent light; launching soon is asked to produce identification. The future indicative is arrested — and charged in the future indictive.
There is also amnesty for survivors. In 1975, Gates and Allen told MITS they had a BASIC for the Altair. Not a line of it existed. They wrote it in eight weeks; Allen loaded it from paper tape in Albuquerque, and it ran — for the first time anywhere — in front of the customer. A false statement in the present tense, zero victims, total canonisation. Nobody calls it fraud; everybody calls it the founding myth of Microsoft. So the two systems ask different questions.
Grammar decides whom it is possible to judge; outcome decides who will be judged.
Before the proof, you are a clown. After the proof, you are obvious.
Between clown and cliché lies the only interesting interval in capitalism — the whole mystery of the recognition market sits there: the stretch where a dream has stopped being a hallucination and has not yet become a fact, where prophecy and fraud share a waiting room and avoid eye contact. I have spent my career inside that interval.
When I first wrote [2012] about fintech, the theses read as a bad habit of confusing the banking system with the internet; then the internet reached the banking system, and the same theses became paragraphs in other people's strategy decks. (It's ok, btw) When I first wrote [2023] about 'Revolut Mafia', it was 'unprooved influencer's bullshit', now it's a noun. When I wrote my notes [2024] across medium and Linkedin arguing that Revolut is bigger than you think — at a moment when its valuation was under pressure — it looked like an excessive fondness for impossible British upside. When I wrote this book [June 2026] arguing that Revolut could become Europe's first trillion-dollar company — the valuation recovered, and the impossible began its slow descent toward a boring row in a spreadsheet. I am not claiming vindication; the book is a bet, and bets settle late.
I am reporting what the interval feels like from the inside: before the licence, "we are building a bank" is metaphysics; after the licence, it is a fact; in between, the founder lives where the present has not yet confirmed the future, but the future is already organising the present.
At seed stage, a founder owns almost nothing: a deck, a voice, a biography, some borrowed credibility, and the future tense. Out of exactly this, occasionally, come 'Revoluts', medicines, rockets, AI models, payment networks. If he succeeds, no one has questions. If he fails, everyone in the room becomes a realist at the same instant, and yesterday's "vision" is re-filed as "misrepresentation" — a manoeuvre performed most confidently by people who enter the hall after the fire and explain how the ladder should have been built.
It sounds like a joke because it is one. It is also, very nearly, a governance framework. A Miranda (or Revolut?) rule for founders would demand neither silence (silence builds nothing), nor permission (permission usually arrives after the people who needed it have left), nor sterilised language (a pitch scrubbed of all imaginative force is not honest; it is dead): mark the tense.
And in return — do not let the accountants of the already-known confiscate the future tense from the people who know how to use it. A society that prosecutes every failed prophecy as fraud will get fewer frauds and fewer prophets; it will become very safe, very correct, very retrospective, and eventually very poor in the only currency that matters before a new thing exists: permission to sound mad for a while.
None of this is a defence of fraud. It is a description of the founder's ontological discomfort — and of where, exactly, the line runs.
The financial system hates deception and runs on rituals of persuasive imagination. It demands the future tense from the founder, the present tense from the auditor, and the past tense from the regulator — then pretends the three can be filed in separate folders. They cannot. They meet in one person, on a stage, holding a clicker.
So state the line precisely. A startup is a legal simulacrum of the future. Fraud is a forged simulacrum of the present. The fraud does not begin when you summon a future that hasn't arrived; it begins when you fabricate evidence that it already has — when you stop exaggerating the gravity of the future and start counterfeiting the mass of the present.
That is the whole rule. The law will ask what was true when you said it; the market will ask whether it became true in time; history will pretend it knew the answer all along. The founder's task is to keep building in the interval before the grammar turns.
So speak carefully. But speak.
Amen.