The Campus Was the Strategy

In 1989, Apple did something that probably did not appear on any strategic planning document as a strategic move.

Engineering was scattered across leased office space in Cupertino. Standard stuff. Cubicles, conference rooms, private offices in the corners. The kind of space that gets built when nobody has asked the question: how does the work actually happen here?

Someone at Apple eventually asked that question. And the answer turned out to matter quite a bit.

Products were not being built in conference rooms. They were being built in the spaces between meetings. The question someone asked walking down the hall. The problem that got solved at someone else’s desk because the person two doors down happened to overhear it. The junior engineer who watched how a senior one handled something difficult and filed it away without knowing she was doing it.

None of that showed up in any workflow diagram. And none of it was surviving the standard office layout.

So Apple commissioned the Infinite Loop Campus with a specific mandate: figure out how engineers actually communicate, and then build the physical environment around that. The project lead later described the core insight simply. The workplace was a communication tool that facilitated the product development process.

Not a place where work happened. A tool through which capability was built.

I have been thinking about that framing a lot lately, because I think most organizations are currently making a version of the mistake Apple was trying to correct. Except now there are two forces at work instead of one, and they are compounding each other in ways that are easy to miss until it is too late.

The first is remote work.

We spent several years arguing about this on the wrong terms. Productivity scores. Collaboration metrics. Whether people work better at home or in the office. Those are real questions, but they are output questions. The more important question is what interactions are no longer happening, and what is being lost in their absence. What is the opportunity cost?

Ambient learning is a terrible phrase for something that actually matters quite a bit. The overheard conversation. The visible decision. The correction delivered in the moment by someone who happened to be standing there. None of that gets scheduled. Most of it does not survive a fully distributed team.

To be fair, remote work is a solvable problem. It is a distribution problem. The interactions can survive if the systems are built to support them: clear accountability, intentional communication structures, leaders who manage outcomes rather than presence. Some organizations figured this out and came through stronger.

A lot of organizations just moved the meeting to Zoom and called it solved.

The second force is artificial intelligence, and it is a different kind of problem.

Remote work separated people physically but left the reason to interact largely intact. You still needed to ask your colleague a question. You still needed a second set of eyes. You still needed someone to tell you your draft was not quite there yet.

AI removes the reason to interact at all. The question gets answered before it gets asked. The draft gets written before anyone has to think through it with someone else. The feedback loop closes without a human in it.

What looks like efficiency is actually the removal of the mechanism by which people learn to do their jobs well. The struggle that produces judgment. The back-and-forth that builds trust. The explanation that goes longer than it needed to because the person giving it was also working something out.

Distributed teams already have fewer of these moments than co-located ones. AI now begins to eliminate what remains. An organization that never built deliberate learning systems for remote work is now absorbing a second disruption that accelerates the same erosion. The gap between what the organization produces today and what it will be capable of three years from now widens quietly, in ways that do not appear on any dashboard until they are very difficult to close.

This is the question I bring to leadership teams navigating both of these shifts: not whether AI is making you more productive today, but whether your organization is still building the people it will need tomorrow.

Apple answered a version of that question in 1989 by studying how their engineers actually learned from each other, and then building the conditions that made it possible at scale. They did not ask how to make engineering faster. They asked how engineering actually worked.

That is a systems question. And it is still the right one.

The leaders who get AI right will identify which human interactions create capability and protect them. They will use AI to clear administrative burden and create more room for the conversations that actually build people. They will treat interaction not as friction to be eliminated, but as infrastructure to be designed.

The leaders who get it wrong will see the productivity numbers move in the right direction for a while.

They will see the other numbers later.

The campus was not a perk. It was the strategy. And the organizations that understood that in 1990, or understand it now, are the ones that tend to build something worth leading.