The Wrong Advice
Most technology commentary about AI is written by people who build software, for people who buy software.
That’s a narrow slice of the economy producing an enormous volume of opinion. The advice isn’t wrong. It’s just written for a different owner.
Different Physics
When a venture-backed technology firm adopts a new platform, risk is already baked into the structure. Capital absorbs failure. Infrastructure is provisional by design.
If a bet goes badly, investors write it off and the team pivots. An acquihire. A Substack essay on failing forward, and then onto the next project. Investors already know the majority of their bets won’t pan out. It’s baked in.
If it works, the technology gets acquired and folded into the next platform. Or the next round of funding earmarks sufficient capital to repair the technical debt as it races to the next tranche of funding.
Rapid experimentation is not recklessness in that context. It’s structurally rational.
Small and mid-sized private businesses operate under different physics.
Ownership stretches across decades, sometimes generations. Growth capital doesn’t come from venture funds, it comes from lenders in the form of personally-guaranteed debt, creating a long tail of responsibility that doesn’t reset between funding rounds. Decisions accumulate. Margins for error are thin. When a system becomes embedded in workflows, training, and institutional memory, migration isn’t a sprint — it’s a distraction measured in months.
If you plan to hold an asset for thirty years, advice optimized for a seven-year exit is misaligned at best. Corrosive at worst.
The Permanent Owner’s Problem
This matters most right now because AI commentary has all but collapsed into a single voice: adopt fast, experiment broadly, be the first mover. The underlying logic is coherent. For a specific kind of company, with a specific ownership structure, operating on a specific timeline.
The permanent owner is working a different problem.
The question isn’t how fast to move. It’s whether you’re taking direction from the right source. Whether you know what makes your business worth choosing in the first place.
Every durable business has an edge that isn’t easily replicated: judgment built over years, a customer relationship that would survive a price war, a craft that signals something a national chain cannot.
That edge compounds slowly. But it can erode quickly. A hype cycle pulls you toward the next thing. You adopt fast, optimize for throughput, and sand down the edges that made you worth choosing — without noticing the erosion until the customer does.
AI doesn’t change what the question is. It raises the volume around it.
For some businesses, the right answer is aggressive adoption. For others, restraint.
What it never is: following a playbook written for someone else’s ownership structure and calling it strategy.
If you use AI to deepen your actual advantage, the technology becomes leverage. If you use it to approximate the throughput of a larger competitor, to look more efficient, to automate what made you distinct, you are spending years of compounded differentiation to win a race you were never built to run.
Play the Long Game
The noise around AI will keep getting louder. Most of it is written for a different owner, with a different exit, and a different definition of winning.
The technology will accelerate whatever direction you’re already heading. Make sure you know which direction that is.
Knowing which game you’re playing is prior to every decision that follows.
