A text file, a few rules: the grid, the margins, the hierarchy, the contrasts. We hand them over to a machine, and it composes. This is the idea circulating through studios this month — design reduced to executable knowledge, recorded in a document an algorithm can follow to the letter. The promise is clear, almost elegant. Yet it revives a much older question: can we really write down what we know about design?
Design as executable knowledge
The shift is real. Files are emerging that describe a visual identity in language machines understand: brand rules, layout logic, typographic scales, contrast levels, review stages, even accessibility requirements. These documents no longer serve to inform a colleague; they serve to direct a tool. The distinction may seem slight. It is decisive.
There is nothing new about codifying design. Brand guidelines and design systems have been doing it for a long time. But those documents still addressed someone: a human who read them, interpreted them and sometimes challenged them. The executable file removes the reader. The rule no longer needs to be understood — only applied.
What the workshop passed on without saying it
Yet design has never been passed on in this way. It is learned in the workshop, through a corrected gesture, in the critique where a teacher says “this does not work” without always being able to explain why. A designer’s knowledge is largely tacit: “we can know more than we can tell,” wrote the philosopher Michael Polanyi. Good letter spacing is not calculated; it is seen.
What we teach a student is not rules — it is how to see. The judgement that knows when a margin breathes and when it suffocates, when a rule deserves to hold and when it needs to bend. A rule can be stated in a line; the exception takes years. That is precisely where the craft lies.
Writing the rule, passing on doubt
Should we therefore distrust these files? Not entirely. Writing them forces us to articulate what is implicit, and that effort is valuable. Setting down the rules of our own work means discovering what we truly believe — an almost educational exercise that lays bare our assumptions and opens them to discussion.
But a file retains the rule, never the doubt. It can say “the heading is forty-six pixels”; it cannot say “pause here, look again”. It knows the norm and overlooks the exception. Yet design lives in the exception — in that suspended moment when the eye decides what no measurement will ever decide on its behalf.
These documents therefore do not signal the end of design education. They mirror it. By recording everything that can be written down, they reveal what resists writing: judgement, hesitation, the eye. And that is exactly the part worth passing on. The machine takes the rule; it leaves us the essential task — learning to see. For those who teach, that may be good news.