I was speaking with someone who might have literally had stars in their eyes. He was glowing with enthusiasm over his newfound ability to develop complex software using an AI Coding tool. He waxed prophetic, predicting the immediate death of the software development world as we know it. This new technology was a revolution, and pretty soon, even my grandmother will be vibe coding her own SAAS tools.
I could not help it; I immediately made a connection in my mind to the Nike Free movement circa 2004. You might remember this, especially if you are a runner. The hype around barefoot running was everywhere, due in large part to a few books that were written on the subject of long-distance runners from indigenous tribes that did remarkably well with minimalist footwear. I remember almost the exact conversation I had about AI coding, but in this situation the starry-eyed young man was talking about his new Vibram FiveFingers shoes. I am sure you remember these: the gloves you would put on your feet and run around in them.
In 2001, Nike designer Tobie Hatfield and other members of Nike’s innovation team visited Stanford and watched coach Vin Lananna’s track athletes running barefoot on the grass. Hatfield asked why. Lananna told him he believed the practice gradually strengthened the athletes’ feet, made them less injury-prone, and therefore allowed them to tolerate more training. Hatfield later described that as Nike’s “aha moment.” See reference article here: https://www.designboom.com/design/tobie-hatfield-nike-free-interview-03-19-2014/
There is an important distinction here. The key concept is gradual. Lananna was not apparently telling his athletes, “Shoes are bad, throw them away.” This also was not a new concept. Barefoot running and minimalist training had existed among serious runners for decades. There was a key principle at play: Don’t let the equipment do so much work that the body loses the ability to do the work itself.
The principle being deployed by the Stanford running team is analogous to a team of expert software developers that start to incorporate new AI coding tools into their practice. The Stanford practice might have been: normal training + carefully dosed barefoot work + soft surface + conditioned elite athletes + coaching supervision.
Unfortunately, Nike’s big moment became the following: Natural running is better. Buy these shoes and run naturally. Now, in Nike’s defense, they had actually built progression into the advertising. They created a free number scale which would help you to slowly migrate from conventional shoes to progression toward barefoot. This is not what happened. The same way no one really buys all the Gatorade drinks for pre-, during-, and post-workouts either.
The result was disastrous. A runner who spent 20 years running in conventional shoes had muscles, tendons, and movement patterns adapted accordingly. Put that running suddenly into a flexible, low-drop shoe, and suddenly you are changing where the stress occurs. The injury data that poured in was not good.
What Stanford had learned is that the human foot benefits from being strong and allowed to move. That knowledge became a conclusion held by the running community that cushioning, support, conventional shoes, and existing training principles are wrong. We have the same phenomenon happening with AI coding tools today.
Experienced engineers use things like abstraction, modularity, tests, source control, code review, architectural boundaries, debugging discipline, and iterative development. AI coding tools observe and automate fragments of those practices. Then the industry starts behaving as though the tool itself has replaced the need for the underlying engineering discipline.
AI can automate parts of software engineering. The danger begins when we mistake those automated parts for software engineering itself. Principle–Artifact Confusion occurs when a sophisticated practice is reduced to one visible component; that component is productized, and eventually the product is mistaken for the principle that made the original practice successful.
When new technology arrives, there is a tendency to abandon basic principles far too quickly. Novelty makes us believe the old rules no longer apply. We invent new terminology, new processes, new organizational structures, and sometimes entirely new philosophies for problems we already know how to solve.
AI does not eliminate the fundamentals of running a good business.
Good data still matters. Clear ownership still matters. Security still matters. Customers still matter. Measuring outcomes still matters. Good management still matters. A bad process does not become a good process simply because an AI agent is performing it.
Before asking what AI can change, make sure the underlying business principles are sound. Technology should amplify good fundamentals, not provide an excuse to abandon them.
