The Stealing From Yourself Philosophy

A founder I have in mind - call him a composite of a pattern I keep meeting - built a consultancy on one thing he did better than almost anyone. A specific kind of analysis. Clients paid a premium for it. Referrals ran on it. For eleven years it was the whole engine.

Then a model learned to do the first eighty per cent of it in nine seconds.

He did not panic. That was the problem. He did what the numbers told him to do, which was to defend the asset. Raise the quality. Add a human-judgement layer on top. Explain to clients why the machine version was inferior. Every one of those moves was locally rational. Each one protected the thing that was still paying for everything.

And each one was a small payment on a debt he had already lost the ability to see.

The asset that made him was the asset now exposing him. He could feel it, and he could not act on it, because acting on it meant turning on the one thing his whole identity had been built to protect.

The asset that made you is the one most exposed

Here is the uncomfortable structure. The thing that produced your success is, by definition, the thing you have optimised hardest, defended longest, and built your identity around. Which makes it the thing you are least able to look at coldly.

Success does not just build an asset. It builds a blind spot in the exact shape of the asset.

Success does not just build an asset. It builds a blind spot in the exact shape of the asset.

This is not a motivation problem. It is not that founders are lazy or unaware. The most alert operator in the market faces the same trap, because the trap is built out of good decisions. Protecting your best margin is a good decision. Serving your most profitable clients is a good decision. Improving the product that made you is a good decision. Right up to the point where the sum of those good decisions is a company that cannot move.

The market has a word for what happens next. It calls it disruption, and it usually tells the story as if the incumbent was stupid. The incumbent was not stupid. The incumbent was rational, and rational was the trap.

Christensen: why the smartest incumbents lose

Clayton Christensen mapped this precisely in The Innovator’s Dilemma (1997). His finding was counterintuitive enough that it took a book to make it land: well-managed companies fail not despite doing everything right, but because they do everything right.

The mechanism is the one my composite founder was living. A new capability arrives that is, at first, worse than the incumbent’s offering on the metrics the best customers care about. It is cheaper, cruder, aimed at the low end or a niche. Every internal signal - margin analysis, customer feedback, resource-allocation logic - tells the incumbent to ignore it and keep improving the core. The disruptive option would cannibalise the profitable business, so declining it looks like discipline.

Then the crude capability improves. It climbs. And by the time it is good enough to take the incumbent’s best customers, the incumbent has no position in it, because every rational quarter said not to build one.

The core insight sits in one sentence: the current success funds the organisation, so protecting it feels rational right up to the moment it becomes fatal. The incentive to cannibalise your own product is negative until it is too late, because cannibalising it means voluntarily shrinking the thing that pays the bills.

Well-managed companies fail not despite doing everything right, but because they do everything right.

That is the dilemma. Not “innovate or die.” The people who tell you to innovate or die have not understood the problem, because the problem is that innovating looks exactly like dying from the inside, at the moment you most need to do it.

The founders who did it anyway

A few operators have broken the pattern on purpose, and the way they did it is instructive because it was never comfortable.

Apple. When Apple launched the iPhone in 2007, the iPod accounted for a large share of the company’s revenue - by most accounts north of forty per cent. The iPhone was, transparently, an iPod killer with a phone attached. Steve Jobs shipped it anyway. Walter Isaacson’s biography records the rule Jobs operated by: “If you don’t cannibalise yourself, someone else will.” Apple chose to retire its own cash engine on its own timing, while it still owned the customer, rather than wait for a competitor to do it on theirs.

Intel. In 1985 Intel was a memory company, and it was losing the memory business to Japanese manufacturers who made DRAM faster and cheaper. Andy Grove tells the story in Only the Paranoid Survive (1996): he turned to Gordon Moore and asked what a new CEO brought in from outside would do. Moore said, without hesitation, he would get them out of memory. Grove’s response was to suggest they walk out the door, come back in, and do it themselves. They exited the business that had built the company, laid off roughly a third of the workforce, and redirected everything to microprocessors. Within a decade Intel was the most valuable semiconductor company in the world.

Netflix. Netflix built a profitable DVD-by-mail business, then deliberately attacked it with streaming, then attacked streaming’s dependence on other studios by building original content. Each transition cannibalised the business the previous one had established. Reed Hastings’ willingness to kill his own working model - more than once - is the reason there is still a Netflix to talk about.

Notice what none of these were. None of them were “embrace change” or “disrupt yourself” as slogans. Each was a specific, painful decision to destroy a functioning, profitable asset on the founder’s own schedule - to capture the value of the disruption while they still controlled it, instead of surrendering it to someone else later.

That is the whole philosophy. Steal from yourself, so no one else gets the chance.

Schumpeter named the weather

Joseph Schumpeter saw the underlying force in 1942. In Capitalism, Socialism and Democracy he called it creative destruction: a process of industrial mutation “that incessantly revolutionizes the economic structure from within, incessantly destroying the old one, incessantly creating a new one.” He called it the essential fact about capitalism.

The word that matters in that sentence is within. Schumpeter was not describing weather that happens to you from outside. He was describing a force that operates from inside the structure - which means the only question is whether the destruction is directed by you or done to you.

Self-cannibalisation is not a defiance of creative destruction. It is the decision to be the one holding the instrument.

The old frame and the real one

Most founder strategy runs on an inherited picture. It is worth naming it directly, because it is the picture that fails.

The shift is from asset to capacity. A specific product, skill, or service is a depreciating position. The capacity to self-disrupt - to see the exposure early, absorb the identity cost, and move before you are forced - is the thing that compounds. It is the only moat that does not erode, because it is made of the willingness to erode everything else.

Why the AI era moves the target from products to skills

Christensen wrote about products and business models. The move you have to make now is to run the same logic one level in - to your own capabilities.

The era changed the target for a concrete reason. Disruption used to threaten a product line, which a founder could hold at arm’s length. Now it threatens skills, which a founder holds as identity. The 2023 study by Eloundou and colleagues at OpenAI, “GPTs are GPTs,” estimated that around eighty per cent of the US workforce could have at least ten per cent of their work tasks affected by large language models, and roughly nineteen per cent could see at least half of their tasks affected. The exposure is not concentrated in low-skill work. Higher-income, higher-education roles showed significant exposure precisely because so much of their value is language and analysis - the exact material these models are built from.

Read that against the Christensen mechanism and the danger becomes personal. The skill you are proudest of is frequently the one most exposed, for the same reason the incumbent’s best product was: it is the most refined, the most language-heavy, the most legible to a model that learned from a decade of people doing it. Your mastery is your exposure.

Your mastery is your exposure. The skill you are proudest of is often the one a model can learn fastest.

And now the trap closes, because a skill is not a product. You cannot look at a skill you have built your name on with the cold eye you would turn on a business unit. It is not in your portfolio. It is in your sense of who you are.

Identity Fusion: the reason you cannot self-disrupt

This is where the strategy problem becomes an identity problem, and why no amount of market analysis fixes it on its own.

In the map of the seven failure patterns that catch high-achieving founders, this one has a name: Identity Fusion. The professional role, or the signature skill, has consumed the person. There is no longer a clean line between “the thing I do” and “who I am.” And when that line is gone, disrupting the skill does not read as a strategic decision. It reads as self-destruction. The nervous system files it under threat, not under strategy.

So the founder defends the exposed skill long past the point where the evidence says to move - not because the analysis is wrong, but because the analysis was never the operative variable. The identity was. This is the exact human version of Christensen’s corporate dilemma: the current success funds the self, so protecting it feels like survival right up to the moment it becomes the thing that sinks you.

You cannot cannibalise an asset your identity is standing on. First you have to get your identity off the asset.

That is what Sovereign Identity means in practice. A sense of self that is independent of any single skill, product, or result. Not indifference to your work - the opposite, a stake in the work that is strong enough to let you change the work. The founder with Sovereign Identity can look at the skill that made him and ask, coldly, whether it is still the right skill to be the best at. The founder fused to that skill cannot even form the question. It feels like betrayal to ask it.

The capacity to steal from yourself is downstream of the capacity to not be your last success.

What this cost me to learn

That is the emotional truth the case studies leave out. Self-cannibalisation does not feel like strategy while you do it. It feels like loss. The founders who manage it are not the ones who feel the loss less. They are the ones who have got their identity far enough off the asset to act through the loss instead of freezing in front of it.

The protocol: what to cannibalise, and when

Philosophy without a mechanism is just a mood. Here is the operational version - a way to find the asset to steal from, and the trigger for when to move.

Step one: run the exposure test on your strongest asset, not your weakest. The instinct is to audit what is struggling. Invert it. Point the analysis at your highest-margin, highest-pride, most-defended capability - the one that makes you money and makes you feel like yourself. That is where the hidden exposure lives, because that is the thing you have never allowed yourself to question.

Step two: ask the three diagnostic questions. For that asset, answer honestly:

1. If a well-resourced competitor, or a capable model, wanted to take this from me, is this the first thing they would target? (High-value, legible, language-heavy assets are the first targets, not the last.)
2. What would I actually lose if this asset no longer required me? (If the honest answer names your sense of importance rather than the company’s survival, the exposure is identity, not economics - and that is more dangerous, not less.)
3. Am I defending this on current evidence, or on the memory of what it used to be worth?

If the asset is the first target, if losing it threatens your identity more than your P&L, and if your defence runs on memory rather than evidence, you are looking at the thing to cannibalise. The discomfort of the questions is the signal, not the noise.

Step three: the timing trigger. You move while the asset is still strong, not when it starts to fail. The entire advantage of stealing from yourself is timing - you capture the value of the transition while you still control it. Waiting for the decline surrenders exactly the thing that made self-disruption worth doing. The correct trigger is not “revenue is falling.” It is “I can see the path by which this becomes obsolete, and I still have the cash flow and the standing to fund what replaces it.” Strength is the window. Weakness is the missed window.

Step four: fund the successor from the incumbent. This is the move the dying incumbents refuse and the surviving ones make. You do not abandon the working asset. You keep its cash flow and deliberately point it at the thing that will replace it - the way Intel funded the microprocessor pivot, the way Netflix funded streaming from DVDs. The winner you are willing to burn becomes the fuel for the winner that follows. That is what makes it discipline rather than recklessness.

Strength is the window. Weakness is the missed window. You steal from yourself while the asset is still worth stealing.

The distinction that decides whether this is even possible

There is a hard limit on all of this, and it is the reason some businesses can self-cannibalise and some structurally cannot.

An asset can be cannibalised and rebuilt. A founder-dependent service cannot. If the entire value of your business is you personally performing the exposed skill, there is nothing to steal from and nothing to rebuild into - you are the asset, and you cannot fund your own replacement because the replacement is also you. This is the difference between an enduring asset and a lifestyle service, and it decides whether self-disruption is available to you at all. The distinction is worth its own analysis, which is why it has one: the enduring asset versus the lifestyle service. Read it as the precondition for everything above. You cannot steal from yourself until you have built something separable from yourself to steal from.

This piece sits alongside others in the same architecture. The Golden Prisoner’s Audit maps where a founder becomes replaceable. The Orchestration Identity addresses the identity a founder installs to move from executor to architect - the exact shift self-disruption demands. Sealed Cognition names how the reasoning behind these calls goes dark when a model sits in the gap where your thinking used to be.

Kasimir Hedstrom | MindMastery