Renting Einstein
On the two clauses nobody read in the strangest employment contract ever signed. Somewhere this morning, a company completed the greatest hire in corporate history.
The new employee reasons like a room full of laureates. He reads a four-hundred-page contract in the time it takes to open it, speaks every language worth speaking, has read approximately everything ever written, never sleeps, never takes a sick day, and costs a fraction of a cent per thought. The board is delighted. The press release writes itself. Although, awkwardly, there is no press release, because every competitor completed the same hire, the same morning, on identical terms.
This is, give or take, the deal every company on earth has just signed. Almost nobody has read the two strange clauses at the bottom of the contract.
Clause one: the new employee has no memory. Not a bad memory. None. Every meeting is your first meeting. He walks in brilliant and blank, and whatever he learned yesterday, the org chart he mastered, the strategy he critiqued so incisively, the correction you gave him at 4pm, is gone by morning. He will greet you tomorrow as a stranger, at genius level. First day, every day.
Clause two: there is no onboarding, and there never will be. He has never met the company. He was not in the room for any decision, any failure, any pivot. He does not know the discount experiment ended badly, that the codebase has a haunted corner nobody touches, that the biggest client threatens to leave every March and never does. He arrives knowing everything about the world and nothing about you.
Watch one week under these clauses.
Monday, strategy. He proposes, with beautiful reasoning, the European expansion. It is genuinely good. It is also the plan the company tried, bled on, and buried two years ago, a fact recorded in a post-mortem he has never seen and will never remember being told. Half the people in the meeting are too new to mention it. The plan is approved a second time. Somewhere in the records room, the post-mortem that killed it sits in a box, filed under a project codename nobody would think to search.
Tuesday, pricing. He asks superb questions. They are the same superb questions he asked last month, which produced the current pricing, which he is now questioning. The analysis is flawless against the premises in the room, and the premises in the room are whatever somebody remembered to carry in. Today it is the deck from the spring review, because it happened to be on top of someone’s drive.
Wednesday, the customer letter. He drafts a renewal note to the company’s largest account: warm, precise, better written than anything the team produces. It quotes a tier that was retired in the spring. The tier is still described, in glowing detail, across nine live documents; the retirement exists only in a decision that never left the meeting where it happened. So the letter is perfect, and wrong in the one way that matters, priced off a belief the company no longer holds. The account manager catches it, this time, then spends the rest of the day rechecking everything else the genius wrote, which was not the productivity curve anyone had in mind. Nobody blames him. Everybody trusts him a little less, for reasons that, fittingly, nobody writes down either.
None of this is his fault, and the fable falls apart if you read it as mockery of the genius. His reasoning is not the weak part. It never was.
By Thursday the company invents a fix: longer meetings. If he forgets everything between sessions, fine. Wheel the archive in at the start of each one. Boxes of it, then shelves, then the whole records room. He reads it all, of course, instantly. And now, with total confidence and perfect citations, he can argue for the buried expansion plan and against it, defend the old pricing and the new, because the archive contains every belief the company has ever held, holds them all at once, and takes no side. More boxes make him better read. They do not make him current.
Friday, an executive says the reasonable thing: “He is the smartest mind that has ever been in this building. He will work out what matters.” And he does work out everything that can be worked out. But which of two recorded plans the company is actually on is not that kind of fact. It was settled in a hallway after the meeting, recorded nowhere, and revised twice since. Genius infers what follows from what it is given. It cannot infer what nobody wrote down. The company keeps buying a smarter reader for a story it never wrote.
The second week, the workarounds begin, because the people are not stupid and the genius is too valuable to give up on. A senior engineer starts keeping a private sheet: two pages of what is currently true, pasted at the top of every meeting with him. It helps, immediately and noticeably. The sheet spreads through the company chat, collects seventeen copies, and by the end of the month the copies disagree, because the sheet is a document too, and keeping it alive is nobody’s job. The company has reinvented its problem one layer down.
Then someone has the properly clever idea: hire him an archivist. A tireless clerk who, before each meeting, fetches from the records room only the boxes that seem relevant. And the clerk is genuinely good at seeming. Ask about strategy, and everything that resembles strategy arrives on the desk. The trouble is that the buried expansion plan resembles strategy more than anything else the company owns; it was argued about for months, so it is the most written-about file in the building, and the clerk carries it in proudly, every time, right on top. The one-page post-mortem that actually killed it, filed under a project codename, resembles nothing much at all. It stays in the room. The clerk fetches by resemblance, and resemblance says nothing about which decision is still alive.
Now, the unfair part of the fable. No company would treat a human genius this way. For a human they would build the boring apparatus without thinking twice: the onboarding week, the handbook, the desk buddy who says “we tried that in 2023, ask Sarah why we stopped,” the walk around the building, the quiet list of things we no longer believe. Induction, in the old HR sense of the word. Companies do this for every human hire without debating it; a junior analyst gets weeks of it before anyone trusts them with a client. The most capable mind ever put on a payroll gets none, at any company, and the omission is so uniform that nobody registers it as a choice. It is unglamorous and it is the entire difference between a consultant and a colleague. The company in the fable spends everything on the genius and nothing on the telling. And since every competitor’s Einstein is identical, same morning, same terms, the only variable left is what each company manages to tell its own Einstein before the meeting ends.
You have noticed by now who the employee is. The genius is the frontier model. The amnesia is the context reset between sessions. The boxes wheeled into the meeting room are the long context window. And the onboarding nobody builds is the knowledge layer: the small, current, adjudicated handbook of what this company believes today, who changed it, and what replaced what. The clerk is similarity search, fetching by resemblance while validity stays invisible to him. The seventeen disagreeing sheets are every private prompt file and pinned note your team already maintains, rotting quietly in parallel. And the handbook works only because one person has the standing to sign what enters it; otherwise it is the records room again, in a smaller box. Capacity is the factor you rent; the handbook is the factor you own. I mapped the full stack in the layers article, and I run my own company on exactly that handbook, which is why my Einstein performs like a partner and not a stranger. If a rented mind ever shows up already knowing which of your decisions are alive, genuinely knowing, in a way you can audit, I will retract every word of this fable in public. Nothing about the contract suggests it. The clauses are structural.
There is a version of Monday that goes differently. Same genius, same amnesia, same terms. But on the desk, before anything else, sits a thin binder: forty claims, each dated, each signed, the dead ones marked dead with their replacements named. He reads it in under a second, the way he reads everything. And the meeting changes shape. The European expansion comes up; he checks the binder, notes it was tried and closed and why, and spends his enormous mind on the question nobody has asked yet instead of the one the company already paid to answer. That morning costs almost nothing to arrange. It is the cheapest thing in this entire story, and almost nobody arranges it.
The greatest hire in history shows up again tomorrow morning. First day, every day. The only question any company controls is what is waiting on his desk.