Published on: July 8, 2026
Ready to accelerate your breakthrough? Send yourself an Un-Robocall™ • Get transcript when logged in
Send Strategic Nudge (30 seconds)Green in-lane · amber a little out · red drift. Every panel is a real commit, byte-identical on recompute. Tap any panel to open its shareable receipt.
Before the argument, the artifact. Run this and read what comes back:
npx thetacog-mcp attest-demo
It returns a signed, recomputable coordinate — where a piece of work landed on a fixed 144-cell lattice, the σ, the lane, the on-chip signature — computed with no model in the loop, byte-identical every time you or anyone else runs it. It does not tell you whether the work was good. It tells you where it stood. Hold that distinction, because the whole post turns on it.
Here is the claim that coordinate is standing behind, stated in its full, attackable form: a résumé is an eval, and it is trying to decide the same undecidable property every AI benchmark is trying to decide. A hiring loop — résumé, interview, reference call, a panel scoring a candidate one-to-five — is a fixed procedure that asks a grader to decide will-this-person-be-good-at-this-role and report a number. That is a non-trivial semantic property of a black box, and Rice's theorem says no procedure decides it — not because interviewers are bad at their jobs, but by construction, the same class of impossibility as the halting problem. The interview isn't a weak measurement of competence. It is a confident measurement of an uncountable thing.
Watch what a credential actually is, because it is the labor market's answer to a cost it could never pay directly. Nobody can decide will-this-stranger-be-good, so the market built proxies: a degree, a former title, a name-brand employer, a warm referral. Each one is a compression of a person's history that a hiring manager treats as a stand-in for the undecidable property they actually care about. The proxy is not the competence. It is a receipt from an institution — this university admitted them, this company retained them — that the manager trusts because re-verifying from scratch on every candidate is too expensive to do. The entire apparatus exists to amortize a verification cost, not to measure anything.
That is why the good ones slip through the cracks and the polished ones sail — the proxy rewards legibility to the proxy, not the work. A brilliant specialist with no brand name reads as a weak signal; a fluent generalist with the right logos reads as strong, and the grader has no decidable way to tell the difference because the property it is grading cannot be decided. This is the same joke running in production at every AI company issuing a safety score: a rigorous-looking count of an uncountable thing. The résumé is a black box's own summary of itself, and determinism was never the alibi it gets treated as — a candidate can be perfectly consistent and perfectly wrong for the lane, and no credential built to measure one dimension measures the other.
The wall does not move. What moves is the question you ask of a person. Instead of will-they-be-good — undecidable, no matter how the panel is staffed — ask did-their-shipped-work-land-inside-the-coordinate-the-role-authorizes. That question has a decidable answer, because it asks nothing about the internal quality of the person or the work. It asks about position: does a compressed representation of what someone actually shipped sit, on a fixed 144-cell lattice, inside the region a compressed representation of what the role declared? That is a joint-compression comparison — gzip, not a model, not a mind reading another mind — and it returns the same bit on every machine, every time. No opinion is in the loop because there was never a slot for one.
Be exact about what this buys and what it does not. It does not certify that a person is talented, diligent, or right. It certifies where their declared intent and their shipped output have landed, over time, on a coordinate a stranger can recompute. A receipt witnesses competence at a location; it does not manufacture it, guarantee it, or grade it. That is a narrower claim than a résumé pretends to make — and it is the only claim in the room that happens to be true, because it is the only one sitting on a predicate that can be decided. A career finally has a floor under it that isn't someone's impression: not "she's a strong engineer," which no one can recompute, but "her last forty shipped units landed in this lane with this σ," which anyone can.
Every hire today is a search followed by a wait. You post, you screen a hundred résumés, you run five rounds, you check references, you extend an offer, you wait out a probation period to find out if the guess was right — months of latency paid whether or not the person was ever going to be fine. That latency is not incidental to hiring. It is hiring; it is the tax the market pays because verifying a stranger's competence was expensive, and the whole ritual is the payment plan.
The receipt collapses the search and the wait into a single step. When a person's shipped work already carries a recomputable coordinate — where it landed, in what lane, with what drift — then finding someone who fits a role is no longer a search through proxies followed by a wait for confirmation. It is a lookup: which coordinates already sit inside this lane? The receipt is not evidence you evaluate after the match. The receipt is the first match — the part that used to cost you a quarter. It settles where the work landed; it does not settle whether this person will thrive on your team, and it never pretends to. Team fit, judgment under ambiguity, the values a good interviewer reads in a room, the early-career or care-work contribution that ships nothing a lattice can see — that stays yours to weigh, and the receipt does not touch it. What it removes is the first tax, the one you paid before any of that human judgment could even begin: reach becomes verify exactly as it does for a model, and procuring the placeable part of a skill, a person's or a bot's, drops to the cost of a cache-line read. The speed of trust is the speed to work — but only for the part trust was ever able to settle. The résumé, the interview, and the probation period were three names for that first tax; remove it, and what is left is the judgment that was always the actual job.
Ronald Coase asked in 1937 why firms exist at all if markets are efficient — why hire someone at a salary instead of contracting each task to whoever on the open market can do it best. His answer was transaction cost: finding a counterparty, verifying they can do the work, negotiating and enforcing the deal, all of it costs something per transaction, and below a certain task size that cost beats just keeping someone on staff. A salary is a bulk-discount on re-verification. You hire a generalist and keep them because re-checking a known quantity inside your walls is cheaper than re-verifying a stranger in the market every single time a task appears. The firm's boundary is drawn exactly at the line where verification stopped being worth it. The book names this directly:
The résumé asks a stranger to certify a semantic property of a person — was this employee good — and that is the same undecidable question already named for a model, asked of a human instead. No non-trivial semantic property of a black box can be decided from its outputs alone, and a résumé is a black box's own summary of itself.
Now run the receipt through the individual's slot, not just the firm's. If a stranger's shipped work can be verified at the cost of a hash lookup, the transaction cost Coase identified does not shrink for the worker — it goes to zero, and the boundary the firm drew around that worker has nothing left to hold it in place. The salary was never only pay; it was a subsidy on the cost of not re-verifying you. Remove the cost, and the incentive flips hard. A career inside a firm rewards generalizing, because being safely re-deployable across the org's tasks is what makes you cheap to keep. A market that verifies a stranger for the price of a lookup does not need anyone to be safely generalizable. It needs them findable — and findable is sharpest at a point, not smeared across a surface. This is the same wall as Section B, standing in the doorway of your own job description.
Here is what the swap gives back to a person, and it is more than efficiency. Three things that were impossible when competence was illegible become ordinary the moment it is placeable.
Infinite specialization stops being a risk and becomes a floor. Today, narrowing too far is dangerous: occupy a niche so specific that no credential describes it and you become invisible to the market — no title, no legible signal, no way for anyone to route work to you. So people hedge into generality and dull their own edge. When the pixel itself is legible — a coordinate anyone can recompute — you can occupy a one-pixel niche safely, because the market can find and price exactly that pixel. The narrowest specialist becomes the most findable, not the least. Depth stops being a bet against your own employability.
The growth path turns into a navigable gradient. Your coordinate has neighbors. The adjacent cells on the lattice — the lanes you are almost in-lane on, the ones a small stretch of your shipped work already brushes — are your visible next rung. Earning potential stops being a fog you feel your way through with a career coach and becomes a gradient you can read: here is where you stand, here is the coordinate one step out, here is what landing there is worth. You can see the ladder instead of being told it exists.
And there is a dignity floor under all of it. Everyone lands somewhere. Everyone has a pixel where their Time-on-Target gives them real standing — a coordinate they occupy that a stranger can verify and therefore cannot wave away. That is not the same as everyone being excellent; excellence stays undecidable and the receipt never claims otherwise. It is something more durable: a place on the map that is yours, provable, that no unearned brand name can outshine and no missing credential can erase.
Read this as the person you actually are, because the wall in Section B retires a different cost for each of you.
If you are looking for work: you stop needing a résumé to vouch for a claim no résumé could ever actually prove. Your drift trajectory — where your declared intent and your shipped output have landed, over time, on a coordinate anyone can recompute — is the thing the market routes work to. It gets sharper the narrower you specialize, not weaker, and it does not care which logos are on your history.
If you hire: you stop paying the search-plus-wait tax. You stop screening a hundred proxies and gambling a probation period on the guess. You look up which coordinates already sit inside the lane you are filling, and the match is the receipt, recomputable before you ever spend a calendar on it — you hire at the speed of the work, not the speed of the interview loop.
If you are a specialist, human or otherwise: you stop being punished for depth. The one-pixel niche that used to make you invisible now makes you the most findable person in it. Your competence is legible at exactly the resolution you earned it, and the adjacent coordinates show you what the next rung is worth before you climb it.
If you run a platform, a marketplace, or a talent network: the scarce thing you are sitting on is about to move. It stops being access to a pool of people — that gets cheap the instant competence is a lookup — and becomes the lane definitions themselves. Which is the honest, uncomfortable part, and the next section says it plainly.
Nothing above asks you to conclude anything on our word — but two things above ask for a warning label, and honesty is the whole product, so here they are.
The limit, restated because it is easy to lose: placement is decidable, correctness is not. A receipt tells you where a person's work landed relative to a lane. It does not tell you the work was good, will be good, or is worth what someone will pay for it — those are semantic properties, Rice's theorem still owns them, and any tool claiming to measure them is selling the same uncountable count the résumé sells. The receipt witnesses competence at a coordinate. It never grades it. If we ever slip and imply otherwise, hold us to this paragraph.
The capture point, said plainly: whoever defines the coordinate lattice sets the standard the market prices against. When competence becomes a lookup, the scarce resource stops being access to skill and becomes the right to define where the pixels are — and that is real power, concentrated, worth naming out loud rather than burying in a footnote. A standard that everyone recomputes against is only as fair as the map it recomputes on. We think the honest move is to make the lattice recomputable and contestable by anyone, not sealed behind a vendor — the receipt is worthless if you cannot check the coordinate system it is drawn on. We would rather you distrust the map and verify it than trust us and take it.
The sources, so you conclude for yourself: Rice's theorem — no program decides a non-trivial semantic property of an arbitrary program — is the standard result behind Sections A and B; the halting problem is the special case most people know. Normalized compression distance — joint gzip compression of declared intent against shipped reality — is the comparator behind Section C (Cilibrasi and Vitányi, Clustering by Compression); it is what makes placement a count instead of an opinion. Ronald Coase, "The Nature of the Firm" (Economica, 1937) is the source behind Section E, and the book's full treatment is where the argument goes past this post.
Three things to actually do, in order of commitment. (1) Run npx thetacog-mcp attest-demo and read the coordinate — a minute, and it is the whole proof. (2) Read the competence-coordinate thesis and the zero-latency market this post extends from AI agents to people. (3) When you want to hold a coordinate of your own, browse the live receipts — each one ringing exactly where the work landed, green in lane, amber bleed, red drift — and hold ours to your own null before you believe a word of it.
The wall does not move for anyone: will-this-person-be-good stays undecidable no matter how many rounds you run or how good the panel is, and no résumé, no credential, no five-star reference average will ever be the thing a market can price. That is not a limitation to route around with better interviews. It is the permanent boundary that tells you which question is even askable — and on the other side of it sits a question that was decidable the whole time, for a model and for a person: not was this good, but where did it land, and can a stranger check.
The labor market that forms first is the one built on the question that can be counted, because a countable event is the only thing capital and coordination have ever known how to route. Everything in this post — the résumé retired, the search-and-wait collapsed to a lookup, the firm's boundary dissolving around the worker at exactly the size where verification stopped being expensive, the specialist findable at a point instead of hiding across a surface, the dignity floor under everyone who lands somewhere — is that one swap, run through a different room. The receipt does not tell you that you are good. It tells you where you stand, and it hands that ground to you instead of to the institution that used to hold it.
The coordinate is on the table. It costs a stranger one second to turn it over. Are you in your pixel, or are you out of it?
npx thetacog-mcp attest-demo. Every load-bearing claim in this post traces back to something you can run or a source you can check — and the one thing it will never claim is that the coordinate means you are good. It means you are placed. That was always the point.