Resistances

This week I had my bypass capability on repository merge rules revoked. It is a capability that allows you to integrate code without going through a formal review. We have spent months pushing for more people to have access to the code — today nearly 80% of the company does — and I have been contributing intensely myself. The result: that capability is now reserved for a small group of developers. I am out. And of course there will be reasons behind it: what happens if everyone can do that; everything could descend into chaos, and so on. But the question is whether these restrictions are introduced because something like that is actually happening — or with a criterion of only those who can also fix their own mistakes without relying on others get this capability — or whether they are a reflex that anticipates it without it ever having occurred.

It is not the first time something like this has happened to me. I have been CPO at a company where, despite my experience as CTO, I could barely touch the repositories because a product person was not supposed to be technical. I have worked at a company where you could not set foot in marketing territory even though technology and product were completely intertwined with it. I have seen entire organisations functioning as systems of departments where people moved tickets from one silo to the next, with nobody ever seeing the whole picture. I have worked intensely to correct that reflex. I have also watched it reproduce itself again.

I have professionally visited dozens of companies. I am fortunate to know many more indirectly, through the window offered by the product management programme I teach. Everything always points to the same reflex. We cannot reduce it to individuals. These are complex dynamics, full of tangled incentives and self-reinforcing inertias.

A dream inscribed in the medium

When Tim Berners-Lee designed what would eventually become the World Wide Web, he faced a design problem that was, at its core, a problem of human organisation. He needed his system to be adoptable by anyone, in any context, without anyone having to ask permission. His solution was radical in its simplicity:

I would have to create a system with common rules that would be acceptable to everyone. That meant as close as possible to no rules at all.

— Weaving the Web, Tim Berners-Lee

It was not just a technical decision. It was a declaration about how people should relate to one another when working with information. Berners-Lee's vision was one of organic, decentralised growth, where anything could connect with anything, where hierarchical classification systems would cease to be the only way to organise knowledge, and where the workings of society would draw closer to the workings of our minds.

He was not a solitary idealist. The people who shaped the different pieces of the internet shared something similar: the intuition that they were inaugurating not just a technology, but a new medium. And a new medium demands new forms of relationship.

It is not just technology

What the web inaugurated was not a more efficient channel for moving information. It was a change in the very nature of who can access, contribute, and decide. It is what Drucker had already intuited decades earlier when he described the entrepreneur as someone who "upsets and disorganizes":

Entrepreneurship rests on a theory of economy and society. The theory sees change as normal and indeed as healthy. And it sees the major task in society — and especially in the economy — as doing something different rather than doing better what is already being done.

— Innovation and Entrepreneurship, Peter Drucker

The web materialised that intuition on an unprecedented scale. Suddenly, the distance between whoever has an idea and whoever can execute it shrank radically. Code could be shared, interfaces could be tested, knowledge could be distributed without intermediaries. Linus Torvalds built an entire operating system with contributors who had never met face to face. Not because the technology allowed it: because the medium demanded it.

Ohno had said it in another way, from a car factory: "People don't go to Toyota to 'work', they go there to 'think'." Grove translated it into the language of management: do not plan like someone designing an assembly line, plan like a fire brigade — form a team capable of responding to the unforeseen.

All of these voices point in the same direction: the new medium needs people who think, who access the material, who cross boundaries. Not people who wait for instructions, guard territories, or process handoffs.

The reflex that never stops

And yet.

In technology, repositories are treated as territory. Whoever has merge access holds power, and that power is protected with processes that seem technical but are often political. The pull request rule, which was born as a quality tool in open-source projects with thousands of unknown contributors, is applied within teams of fifteen people sitting in the same office. The instrument of openness becomes an instrument of control.

In design, user flows are considered the property of the design team. For someone in technology or business to propose a flow is perceived as an invasion, not a contribution. The disciplinary boundary is defended as if product quality depended on each craft working in its own box, when experience shows time and again that the best products are born at the intersection.

In organisations, the tendency to form departments — with their heads, their metrics, their own objectives — remains the dominant model. Autonomous teams with different perspectives, capable of owning a problem from end to end, are still the exception. The norm is the chain: one defines, another designs, another builds, another tests. Each link delivers to the next and disengages. Nobody sees the whole. Nobody owns the outcome.

These resistances are not new. They are the same ones that appeared when the web began to change the rules. The same ones Cagan described with resignation when he pointed out that companies keep Agile's rituals without modifying the engine. The same ones that have been creating tension within organisations for thirty years.

The latest turn of the screw

Now large language models are tightening that same screw once more. When anyone can write code, analyse data, prototype an interface, or draft a technical specification, the boundaries between disciplines blur even further. The space separating whoever has the idea from whoever can execute it — which the web had already compressed — narrows to the point of almost vanishing.

And the resistances repeat themselves with a precision that would be comical if it were not so costly. Who can use AI on the team? Do you need permission? What happens if someone from the business side generates code? What happens if someone from technology designs a screen? The same questions, the same boundaries, the same reflexes of territorial protection.

It is not that people are petty or short-sighted. It is something deeper: organisational inertia has an industrial DNA that refuses to die. Specialisation, the chain of command, control over who touches what — all of that worked in a world where the material was physical, mistakes were expensive, and information was scarce. In a medium where the material is malleable, mistakes are cheap, and information is abundant, that DNA becomes a burden.

Resistance as a symptom

There is something I have learned to see over time. The resistances are not the problem. They are the symptom of a deeper tension: the tension between the way the medium asks us to work and the way we have learned to work. Between what the web inaugurated thirty years ago and what organisations still have not assimilated.

Berners-Lee wanted a system with the fewest rules possible. Not because rules are bad, but because he understood that a decentralised, organic, connected medium cannot function with the hierarchical classification systems we cling to. Every time an organisation locks down a repository, protects a disciplinary territory, or adds a link to the approval chain, it is clinging to a way of working that the very medium in which it operates has been trying to dissolve for decades.

It is not about having no rules, or about everyone doing everything. It is about understanding that every access restriction carries a cost that is rarely accounted for: the cost of the person who had something to contribute and could not. The cost of the connection that was never made. The cost of the team that never got to think together.

Ohno said that people went to Toyota to think, not to work. But for that to happen, people need access to the material they are thinking about. And that — granting access, letting go of control, trusting that the team will know what to do with what it has — remains the hardest resistance to overcome. Not because we do not know it must be done. But because every time we try, the old reflex asks us to close the door we have just opened.