I write a philosophy in public, and the responses arrive. People say it resonates. They quote it back to me. They tell me it changed how they think. I have learned to treat almost all of it as noise. Not because it is insincere, but because it costs nothing to produce, and a signal that costs nothing carries almost nothing.
This is the uncomfortable thing about putting ideas into the open. The ideas attract agreement, and agreement is the least informative response there is. A person who agrees with me has told me about their taste, not about the idea. Worse, most agreement is a kind of wearing: the reader picks up the vocabulary, repeats the moves, signals membership in whatever the ideas seem to define. That is mimicry, and mimicry is what ideas attract most easily, because copying the surface of a thought is so much cheaper than thinking it.
So I want a harder test. Not "who agrees," but "who actually engaged," where engagement has to cost something to count. I can find only two gates strict enough to be worth anything, and almost nobody clears both.
The first gate is word that extends. Restating an idea, however eloquently, proves only that you can restate it. The test is whether you took the idea somewhere its author did not: pushed it into a domain it had not reached, derived a consequence the author missed, built a mechanism the original only gestured at. Extension is costly because it requires understanding the idea well enough to operate it, not just admire it. A reader who extends has stopped being a reader. The echo tells you the idea was heard. The extension tells you it was understood.
The second gate is deed that costs. Ideas about how to live, how to build, how to treat people all imply actions, and the actions are where the belief becomes legible. Anyone can write that you should bet on people before the world does; the question is whether you wired money to someone unproven. Anyone can theorize that the right way to help is to make yourself unnecessary; the question is whether you have actually let someone walk away. Skin in the game is the old name for the gate, but the sharp version is narrower: the deed has to spend in the direction the idea points even where that direction runs against your own interest. Self-interested cost can be bought with status and counterfeited with it. Cost paid against yourself cannot.
The gates are strict on purpose, and most candidates die at one of them. The theorist dies at the second: she writes the better version of the idea, extends it cleanly, and never spends a dollar making it real, so the philosophy stays a literature. The patron dies at the first: he funds people generously and at genuine cost, but his money moves on instinct he never connected to any idea, so there is nothing to extend, only to admire. The largest group, the people who quote the idea and feel moved by it, die at both gates at once, having neither moved the thought nor paid for it. The case that survives writes something that carries the idea forward and then goes and does the thing the idea demands, with resources that were theirs to lose. Going in, my prior is that this case does not exist for any given philosophy. The search is for a disconfirmation.
For my own philosophy (own your own mind rather than rent it; compound knowledge in the open; bet on people before they are obvious; build things that graduate the people who use them) I went looking, adversarially, for someone outside my own project who cleared both gates in public. I expected no one. I found one.
There is an essay called "The Future of Loving Grace," published at a personal blog under the name Andy. The title echoes a familiar phrase, the one Brautigan used for a world watched over by machines of loving grace and Dario Amodei borrowed for his case that powerful AI could go well. The essay does not dwell on that lineage. It spends itself on the part the optimistic visions leave abstract: the actual institutions that would raise and sustain a person across a life.
It imagines for-profit companies (it calls the archetype "Omega School") that act as parenting copilot, tutor, and life coach, competing in a market whose pressure is meant to align them with the flourishing of the people they serve. A child is "enrolled" and given a small equity stake in the company that raises him, so the company's upside and the child's are tied together. There is a survival floor that keeps a person solvent while they grow, and the stated goal is graduated independence: the relationship is built to end. The author calls the ending a "joyous parting," the point where the coaching has succeeded precisely because the person no longer needs it and is sent off to find something better. Gifts back to the company, in the essay's words, "demonstrate gratitude for promises kept," not the discharge of an obligation.
I did not write any of that, and almost all of it sits one step from things I did write. A system whose success condition is that the user outgrows it and leaves is the exact objective I argue an honest AI should optimize for. The equity stake that ties a person to whoever invested in them early, and the gratitude that settles by being passed on rather than paid back, is the argument I had just finished making about the unpaid debts every successful life carries. The premise the essay states outright, that "morality is asymptotically solved by market mechanisms," is a position I hold about why business runs on a physics rather than a morality play. He reached it from a different direction, in different examples, aimed at a different question. That is extension, not echo. He cleared the first gate without my help and possibly without my knowledge.
The second gate is where most thinkers stop, and where this case turns unusual. The same person runs a grant program called &U. It is not a manifesto about generosity. It is a ledger. In 2025 it moved more than a hundred and eighty thousand dollars to founders and ambitious people, in amounts from a hundred dollars to fifty thousand, with named recipients and dates.
Read the terms against the philosophy. The headline is "a completely free $50k with no strings attached at all," to people chosen for "exceptional grit and an extremely high rate of learning," and explicitly, "you do not need any product ideas or a résumé." That is the bet-on-people-before-the-world-prices-them claim, executed with real money on exactly the people who have no résumé to price. Recipients are asked only to "someday, pay it forward," which routes the gratitude onward rather than back, the same shape as a gift that settles a debt by extending it. The application form is treated as the product: applicants report that filling it out clarified their own ambitions whether or not they were funded, which is the thing a demanding form does, standing in as an evaluator's taste compressed into a structure you measure yourself against. And the grantor inverts the usual posture of patronage by thanking the applicants for showing up at all.
Every one of those spends against the spender's own interest, which is the gate's hard version: no strings retains no upside; the no-résumé rule accepts unhedged risk; building the parting into the design gives up the customer on purpose; thanking the applicant gives away the leverage a grantor is handed for free. Even the winding-down counts. The author has said grants are no longer his main focus in 2026, and a thing built to make itself unnecessary is allowed to end. The deed gate is not cleared by intending to be generous. It is cleared by the bank statement, and the bank statement is public.
Here is where I have to be careful, because the temptation is to claim too much. I cannot show that this person got any of it from me, and I cannot show that he didn't. The direction of the arrow is genuinely unknown to me: he may have read my work and run with it, arrived at the same place on his own, or gotten there first. The result barely depends on which.
If he converged on his own, then the commitments we share (graduate the user, bet early, let gratitude run forward, trust the aligned market, build the exit into the design) are probably not my house style but features of the territory, the pass that different climbers find through the same mountains because the mountains are shaped that way. If instead he took the ideas from somewhere and ran, then they were operable enough to extend and actionable enough to spend on, which is a different proof of the same point. Either way the test returns what agreement never could. An idea echoed back at you is only evidence about your reach. An idea that a stranger extends in writing and pays for in public is evidence about the idea.
So the search did not find me a disciple, which is what it was quietly hoping for and what would have proven the least. It found an instance: one person, outside my project, someone I do not control and cannot claim, who took the same ideas seriously enough to carry them forward in word and expensively enough to enact them in deed. That is a stricter result than agreement and a better one. Agreement would have told me the ideas are likable. The instance tells me they might be true.
It is a sample of one, and a winding-down one, and a speculative essay paired with a closing checkbook. I am not going to inflate it into a movement. But the whole point of two gates that strict was that passing them would mean something, and one real pass outweighs a thousand resonances. I went looking for evidence that the philosophy was more than my own voice in a closed room. I found someone who had written its next chapter and paid for its first.