Reflections from Mesh Lab Episode 2: Dismantle First, Then Build

We do not have a human skills problem. We have a condition problem.

And until we are honest about that distinction, every program we build to develop curiosity, adaptability, and judgment will fail for the same reason the last ones did.

Watch a Child in a Sandbox

Here is the unavoidable truth: human capabilities are not missing. They were never absent.

Think about a group of five and six-year-olds in a sandbox. No instructions. No one is telling them what to build or how to do it. One kid starts digging. Another has a different idea and says so. They negotiate, without any adult facilitation, until they land on something both can live with. A third one tries something, watches it collapse, and without a single moment of shame just tries again. They are reading each other constantly, adjusting, creating, and persevering. 

Then school starts. Really starts. And over the next twelve years, the sandbox gets taken away.

Compliance-based schooling rewards the student who stays quiet, follows the lesson, and produces the expected output. Hierarchical organizations reward employees who execute predictably and work to please their boss. Neither asks what human potential is already in the room. Both spend years removing it. The system did not fail to develop human capability. It was designed to suppress it. Compliance requires predictability. Predictability requires conformity. Conformity requires that curiosity, creative risk, and the willingness to try and fail and try again be conditioned out early and consistently.

AI did not create this problem. It made the cost undeniable.

When a machine can generate the outputs your system was built to produce, content knowledge, standardized answers, predictable execution, the credential that proves you checked the boxes, what exactly is left to justify the system? What remains is exactly what the system spent years conditioning out. The capacity to navigate ambiguity. To build trust across differences. To ask a question nobody has thought to ask yet. To make a judgment call and own the consequences. These are not new skills. They are what was alive in that sandbox before the bell rang.

The Program Is Not the Problem

The instinct is to respond with more programs. Resilience training. Empathy modules. Leadership workshops on psychological safety. We keep building them. We keep wondering why they do not move the needle.

They do not work because the delivery mechanism is not the problem. The conditions are.

You cannot develop curiosity in an environment where the right answer is predetermined. You cannot build adaptability in a system where deviation is penalized. You cannot grow judgment when every decision gets escalated upward, and the message sent to people at every level is that their thinking is not trusted. And you cannot develop the relational intelligence that AI cannot replicate by putting people in front of screens for the majority of their learning and working lives.

Think about what flat communication actually costs. A team meeting with an agenda packed with status updates, leaving no room for anyone to think out loud. A new employee whose questions get redirected to a document rather than a person who can sense what they actually need. A performance conversation where the form drives the dialogue and the human in the chair is secondary to the process. A classroom where the lesson plan leaves no space for questions and discovery. These are systems that signal, at every turn, that human judgment is not the point. That thinking is inefficient. The role of the person is to move through the process, not to bring themselves to it. These are not edge cases. These are the operating conditions of most organizations and most schools. And we are surprised that judgment, trust, and relational intelligence are in short supply.

The environment is the curriculum. What we build around people teaches them how to be. And right now, most of what we build teaches people to perform, defer, and wait for direction. The sandbox is not coming back through a workshop. It requires a fundamentally different design logic.

You Cannot Give What You Never Had

Here is what makes conditions so difficult to change, even when leaders genuinely want to change them.

Most of the people being asked to redesign conditions were themselves products of the suppressive system. They learned that compliance earns safety. That certainty signals strength and vulnerability signals weakness. The path forward is to please your boss, not challenge the norm. And now we are asking those same people to build environments of psychological safety, trust, and autonomy for others.

You cannot give what you never had. You cannot model vulnerability if you were punished for it. You cannot create permission for failure if you were never given that permission yourself.

This is why conditions are a leadership problem before they are a design problem. Someone has to go first. Someone has to be willing to say what they do not know. To share the decision that did not go as planned. To ask what you think before offering the answer. To choose the discomfort of honesty over the false safety of control.

That is not soft leadership; it is the most strategic action available. Because the moment a leader models genuine vulnerability, they create permission for everyone around them to do the same. The sandbox does not come back because someone installed a program. It comes back because a person in authority signaled that it was finally safe to show up human.

Unmeasured Is Not Unmeasurable

Redesigning conditions demands accountability. And this is where the hardest conversation lives.

The argument I hear constantly is that human capabilities are too soft to measure. That belonging, agency, curiosity, and judgment cannot be tracked with the same rigor as test scores and revenue targets. That argument is a convenient excuse for systems that do not want to be held responsible for developing the whole human.

Unmeasured is not the same as unmeasurable. We can observe reasoning in real time. We can document how someone navigates a problem without a clear answer. We can track whether people in our systems are growing more adaptive, more confident, more capable of taking ownership over time. The tools exist. What is missing is not the methodology. It is the will to build accountability structures around what actually matters.

AI can generate the answer. It cannot stand behind a decision and explain the values that shaped it. It cannot be held accountable for consequences. The gap between generating output and owning judgment is where the human advantage lives. That gap is visible. That gap is measurable. Making it visible is the work.

Until we measure human capability with the same rigor we apply to everything else, we will not fund it. Until we fund it, we will not build it. Until we build it, we will keep producing compliant outputs in a world that no longer needs them.

Dismantle First, Then Build

We are two sessions into this twelve-month series. And the question I am sitting with is the one the whole conversation has been circling.

Not what do we add? What do we dismantle?

What structures are we maintaining that actively suppress the human capability we say we want? What would it mean to measure what actually matters and build real accountability around it?

The sandbox is still there. It never left. The six-year-old who negotiated, created, failed, and tried again did not lose those capabilities. They were conditioned out by systems built for a world that no longer exists.

Dismantle those systems first. Then build the conditions where the sandbox never had to be taken away.

The future is human. But only if we build it.