AI StrategistRich Schefren · Strategic Profits

Field note

From 2006 to 2026

The old advice was to remove yourself, and it was right. What reversed is not the goal. It is which of the two things a business runs on can be carried without you in the room.

Where this comes from

The 2006 position is the Internet Business Manifesto, published 22 June 2006 and downloaded more than five million times. The 2026 position is The A.I. Business Manifesto. Both are the same author's, twenty years apart, and this note sets the two side by side.

Not published: Nothing. Both documents are public and free.

What changed between 2006 and 2026 Two panels. In 2006 execution was the scarce input, so the advice was to standardise and scale it, while judgment could not be carried and so was removed from the workflow. In 2026 execution is abundant and can be rented, while judgment can now be captured and multiplied. The solid box in each panel marks the scarce input, and it moves from execution to judgment. 2006 The blueprint that built real companies EXECUTION Scarce. Costly. Only scales if it is standardised first So: systemise it JUDGMENT Cannot be carried. It only exists while you are present So: take it out of the work Stop being the bottleneck. Build the systems. Remove yourself. WHAT CHANGED Execution became abundant. Judgment became capturable. Both facts moved, in opposite directions. The advice follows the facts. 2026 The same goal, the opposite instruction EXECUTION Abundant. Competent, fast, and available to everyone So: rent it JUDGMENT Now the scarce input, and for the first time carryable So: capture it, then multiply it
The solid box marks the scarce input. It moves. That movement is the whole argument, and everything else follows from it.

The same diagram in words

20062026
ExecutionScarce and costly. Scales only if standardisedAbundant. Competent, fast, available to everyone
What to do with itSystemise itRent it
JudgmentCannot be carried. Exists only while you are presentThe scarce input, and for the first time carryable
What to do with itEngineer it out of the workflowCapture it, then multiply it
The instructionRemove yourselfMultiply yourself
WhyYou were the part that could not be standardisedYou are the part that cannot be rented

Why the 2006 advice was right

It is worth being clear that this is not a correction of a mistake, because the usual version of this story is that someone got it wrong and has now seen sense.

In 2006 a business could scale exactly what it could standardise. That was not an opinion, it was the mechanics of the thing. Anything that depended on a specific person's call could not be handed to more people, could not run in more places, and put a ceiling on the whole operation at the size of one person's attention.

Given that, there was one available move. Simplify the work. Standardise the decisions. Turn what you could into rules, and take what you could not out of the workflow. The parts of the business that depended on your particular judgment were, correctly, treated as defects to be engineered around.

That advice gave owners their time back and built real companies. It was right.

What actually moved

Two facts, in opposite directions, and both had to move for anything to change.

Execution stopped being scarce. Competent work of most kinds is now fast and cheap. That is a genuine gain and it is also why systemising stopped being an advantage: your competitors got the same capability on the same terms in the same month. Standardisation used to be a moat because it was hard. It is now table stakes because it is not.

Judgment stopped being untransferable. Not because anything mystical was solved, but because the calls can now be caught at the moment they are made and applied later, which is a logistics change rather than a metaphysical one. Apprenticeship always transferred judgment. What it required was ten years of two people standing next to each other waiting for rare cases to occur.

Why the prescription flips

Advice is a function of what is scarce and what is transferable. Both inputs reversed, so the output reversed.

When the untransferable thing is also the scarce thing, you have to design around it, which is what remove-yourself was. When the scarce thing becomes transferable, designing around it means designing your only remaining advantage out of the business.

In 2006 you were the part that could not be standardised. In 2026 you are the part that cannot be rented. Same person, opposite instruction.

Which also explains why this reads as contradictory to people who took the first piece of advice seriously, and why it is not. The goal never changed. The goal was always a business that produces value beyond what one person can personally do in a day. The only question was ever which component could be sent out into the business without you attached to it, and the answer to that question changed.

The part that is uncomfortable

Twenty years of good practice trained a specific instinct: when something depends on your judgment, that is a problem to be removed.

That instinct is now pointed at the wrong target, and it is strongest in the operators who followed the advice best. The businesses that did the most thorough job of standardising themselves are the ones with the least captured and the most systematised, which is the exact position from which the current shift is hardest to exploit.

There is no comfort to offer about that beyond the obvious: the judgment is still in the owner's head, it did not go anywhere, and the capture can start on any Tuesday.

Frequently asked

What changed between 2006 and 2026?

Which of the two things a business runs on can be transmitted without the owner. In 2006 execution was scarce and could only scale if it was standardised, while judgment could not be carried at all, so the correct move was to systemise the first and engineer the second out of the workflow. Execution is now abundant and judgment can be captured, so both halves of the instruction reverse.

Was the 2006 advice wrong?

No. It was correct for the constraint that existed and it built real companies. Advice is a function of what is scarce and what is transferable, and both of those have moved. A prescription that reverses when the facts reverse was a good prescription.

Does this mean stop building systems?

No. Systems still matter and execution still has to run. What changed is that systemising is no longer where advantage comes from, because competent execution is available to your competitors on the same terms and at the same speed. Advantage moves to the part that is not available to them.

What does 'multiply yourself' actually mean?

That the calls behind your best work get captured and applied in more places than you can personally be. Not that you work more, and not that you appear in more places. The judgment operates in parts of the business you are not in.

Why is this the moment rather than five years ago?

Because both conditions had to hold at once. Execution had to become good enough that standardising it stopped being an advantage, and capture had to become practical enough that judgment could be held outside a person. Either alone changes nothing.

Where this sits

What judgment becomes once it can earn while you are not applying it is at judgment capital. Why the transfer used to cost a decade is at apprenticeship without the decade. The category is defined at what Imprinted AI is.

Related

Keep going

Twenty years apart, two documents, one reversal. The second one is the argument this whole site is downstream of.

The full case is The A.I. Business Manifesto. About 23,000 words, free to read on the page, no gate in front of it. If you would rather have the short version, the same page will send you the three fixes and the PDF.

Read The A.I. Business Manifesto

Free either way. Reading it costs nothing and asks nothing.

Last updated: 28 July 2026