Feeling the materials
Design leaders work at distance from the things their teams make. That is the nature of the job. You shape the conditions in which work happens, and you trust that the output reflects those conditions. The higher up in an organisation you go, the more layers sit between your judgment and the thing that ships.
AI adds more of them. Our teams are building faster, through more abstracted tooling, with less visible craft in the process. The gap between your instincts and the finished product is wider than it has ever been.
That gap is where things can go wrong in one of two ways.
The first is obsolescence. This is what happens to leaders who lose contact with the material entirely. Who stop making things. Who only ever commission and review and sign off. It happens gradually and it is easy to rationalise. You are busy, and your job is strategy now. But instinct is not something you can keep without using it. At some point, the process is telling you what is good and the metrics are telling you what is working, and before you know it you have no independent means of interrogating either.
You cannot feel the difference between something that is right and something that just looked right. When the landscape shifts, there is nothing to fall back on. You are a manager of managers who has literally lost touch with the output you were managing towards.
The second is dependency, and this failure mode is probably the more common of the two. These are not leaders who have stopped making things. They are engaged, they are curious, they are shipping. But they are doing it entirely through layers and tools and processes they do not understand and cannot interrogate. The tools are doing the discernment. They are approving outputs they can’t evaluate, making calls based on velocity and volume because that is what is legible to them.
Their judgment is rented. When the tool changes, or produces something wrong, they cannot see it. They have the confidence of someone who is still in the game, and the blindspot of someone who has unwittingly left it.
Neither failure mode is new. Obsolescence and dependency have long been failure modes for senior leaders. What AI does is accelerate both.
Good architects do not build walls, but they always handle the materials. They know what concrete does in different temperatures. They know what a threshold feels like when it is wrong by 20mm. They build scale models. They keep sample libraries. They visit buildings under construction, and they live in finished ones with professional attention. This is how judgment stays sharp across a career that takes you further and further from the making. And the best of them do not confine it to working hours. They sketch buildings that will never be commissioned. They notice how a room projects light on a holiday they are meant to be relaxing on. The appetite does not switch off when the project does.
The web has always had a version of this. Before product organisations and roadmaps, people built things for the internet because they wanted to. Personal sites, strange tools, experiments that went nowhere. That culture still exists, but has been drowned out by OKRs and podcasts venerating venture capital.
This is where AI gets interesting. In the industrial context it is a productivity layer, and the conversation stays stuck there. In this other mode, personal, exploratory, free of organisational stakes, it is a bigger set of hand tools. The same capability that lets it manufacture output you cannot interrogate also lets you reach further into a material than you could on your own. You can build the strange thing that was previously beyond you. You can follow a curiosity past the point where your own skills would have stopped you. This is the version of AI that feeds judgment rather than renting it. You are still the one deciding what is worth making and whether the thing in front of you is any good. The tool extends your reach without taking over the discernment, because nothing is at stake except whether you wanted it to exist and your ability to guide an industry that needs your taste and judgement more than ever.
So, keep making things. Maintain a live relationship with the internet as a place you build stuff outside any organisational context. Not to stay technical, but to keep feeling the difference between what is right and good, versus what is bad and wrong.
Taste and judgment need feeding. This is how we feed it.