The Trap of Precedent

We built the modern organization around a simple idea: knowledge is the asset. The person who had seen more, done more, studied more — that person was valuable. We created credentials to certify it, hierarchy to deploy it, and entire management systems to make sure it flowed in the right direction.

AI just showed us how fragile that idea always was.

A lot of what we called knowledge was pattern-matching dressed up as wisdom. The expert had practiced answers, not necessarily better ones. The system — the credentials, the deference, the hierarchy — was mostly built to manage our collective anxiety about not knowing, not to shrink the not-knowing itself.

A smart person with AI and good questions can now out-produce a twenty-year veteran in the same domain. Not in every case. But in enough cases, and across enough domains, that we cannot ignore it anymore. The gap between knowing and performing knowing has been exposed. That changes everything.

The Comfort of Precedent

Precedent feels like wisdom. It presents itself as experience, as having seen this before, as pattern recognition earned over time. Sometimes it is those things. But most of the time it is something simpler: it is comfortable. Precedent tells you what to do so you do not have to sit with not knowing.

That comfort is the trap.

Not because experience is worthless. It is not. But because precedent becomes a ceiling. It quietly steers you toward what has worked before and away from what has not been tried. It draws invisible boundaries around what is possible. When the world changes slowly, that is manageable. When the world changes as fast as it is changing now, it is a liability.

Look at how most organizations make decisions. They hire for credentials — certifications of past performance. They promote for tenure — rewards for familiarity. They defer to the most experienced voice in the room. The entire structure is optimized to extract value from what has already been figured out.

That worked for a long time. It does not work now.

What AI Actually Exposed

People talk about AI as a productivity tool. That is true, but it misses the point.

The more important thing AI did was expose the assumptions we had about knowledge — who has it, what it is worth, and how organizations should be built around it. When any competent person with AI access can generate expert-level output across almost any domain, the question of who knows something becomes far less important than the question of what we are actually trying to figure out.

AI has no ego. It has no career to protect. It has no investment in the way things used to work. It does not slow down to manage its own uncertainty. It just tries things.

Working with that kind of tool changes how people think — if they let it. It makes experimentation cheap. It compresses feedback loops. It models a different relationship to being wrong.

The people who are genuinely thriving with AI are not the ones with the most experience. They are the ones without deep attachment to how things used to work. No priors to protect. No assumptions to defend. They are just exploring. And that turns out to be enormously powerful.

Collaboration Without Precedent

The best collaboration I have seen does not happen between established experts working in their lanes. It happens when new people do new things in new ways — without the accumulated weight of how it is supposed to work pressing down on every decision.

We have been taught the opposite. Strong teams, we were told, are built from complementary expertise. The designer, the developer, the strategist — each bringing their piece, fitting together like a puzzle. The problem is that this model has a hidden flaw: the interfaces between specialists are defined by their prior assumptions. The designer assumes certain constraints. The developer assumes certain requirements. The strategist assumes certain markets. Nobody makes those assumptions explicit. They just quietly limit what the team can imagine.

Take away the specialization — or at least its weight — and something different happens. People stop defending territory. They stop managing the gap between what they know and what they do not know. They get into it together. The questions get better. The thinking gets stranger. The solutions stop looking like everything that came before.

AI accelerates this by democratizing capability. When everyone on a team can write, code, analyze, and design, the rationale for traditional team structure dissolves. What is left is the real question: how do these people actually think together? And the answer is that thinking together works best when no one is performing a confidence they do not have.

Wrong as a Feature

Every great team I have seen has a specific moment when it becomes genuinely collaborative. It is the moment when being wrong stops feeling like a threat and starts feeling like progress.

You cannot schedule that moment. You cannot workshop your way to it. But you can build conditions for it. Model it yourself — be visibly wrong and visibly fine with it. Reward the question more than the answer. Structure the work so that trying and failing costs less than not trying. Hire people who get curious when something breaks their mental model instead of people who shut down.

The environment is the leader's job. What emerges from it is not something you can control. It has to happen on its own — through enough small moments of shared uncertainty that people stop managing their image and start actually thinking together.

When it happens, you know it. The quality of information flowing through the team changes completely. People share half-formed ideas because the half-formed idea might be the real one. The obvious question finally gets asked — and it turns out to be the right question. The thing that was not supposed to work gets tried, and it works.

AI makes this culture more viable than it has ever been. Test a wrong idea in an hour, not a quarter. The feedback loop tightens. The culture that treats wrongness as interesting gets reinforced faster, and it deepens faster. Being wrong becomes a feature, not a bug.

Where Specialization Goes

None of this means expertise becomes irrelevant. It means the location of expertise shifts.

Specialization moves from generation to application. Anyone can produce the analysis, the code, the document. Knowing which problem deserves that effort — and what to do with the result — is still hard. That judgment is built through years of lived experience in a specific domain. It compounds. It does not get commoditized.

The specialist of the future is not someone who can do something others cannot. It is someone who understands something others do not — a specific context, a set of edge cases, a domain where the wrong call has real consequences. Their value is not in the execution. It is in knowing where to aim.

This makes real domain experience more valuable, not less. The person who has spent years in logistics, healthcare, or financial operations — and who also knows how to apply AI — holds a compounding advantage. AI amplifies domain knowledge. It does not replace it.

What to Build

The organizations that figure this out will not win because they have better tools. Everyone has access to the same tools. They will win because they built the right culture around those tools — one where precedent is a reference point and not a ceiling, where uncertainty is shared and not managed in private, where being wrong is the beginning of something and not the end.

That culture requires small teams. Coordination overhead kills it. Every person added expands the surface area for assumption-collision and delay. The teams that move are the ones where everyone sees the whole picture and trusts each other enough to act without waiting for approval.

It requires different hiring. Less emphasis on credentials that certify what someone has already done. More emphasis on how someone thinks when they do not know the answer — whether they get curious or defensive, whether their questions improve over time, whether they can stay comfortable in uncertainty and still move.

And it requires a different kind of leadership. The leader who has the most experience and the most practiced answers is no longer automatically the most valuable person in the room. The most valuable person is the one who creates conditions for the team to think together in ways none of them could manage alone.

What emerges from that cannot be planned. That is the whole point.

The trap of precedent is comfortable. That is what makes it a trap.