Deliberate Motion

We don't really think in clean systems or linear sequences. Thought arrives in fragments. One idea interrupts another. Something that felt clear a few minutes ago starts to feel wrong once you see it taking shape. You follow one thread, double back, throw part of it away, then notice something useful in what looked like a dead end.

That has always been one of the things I loved about design. Long before language models, I worked this way. I would start with a faint sense of how something should feel, give it form, look at it, change it, undo it and start again. Sometimes I would spend hours on a direction before realizing the premise itself was wrong. Some of the better work came from being willing to abandon what I had already made rather than trying to rescue it simply because I had invested time in it.

Making gave me something to think against. Once I could see an idea taking shape, I could react to it. I could notice what needed to be simplified, where I had forced something, or where there was more to explore. Very often, I only understood what I was trying to make after I had already made some version of it and had something concrete in front of me.

The tools we have now compress that loop considerably. An incomplete thought can become something tangible almost immediately. I can move between words, interface and code while an idea is still forming, then respond to what is actually there instead of trying to preserve every part of it mentally until I have enough clarity to begin.

There is something genuinely useful in that. Language models are often surprisingly good at recovering intention from fragments, shorthand and badly formed sentences. You can give them something half articulated and get back a version coherent enough to examine. For someone who already works by externalizing ideas and reacting to them, that can feel like a natural extension of an existing practice.

But the convenience introduces another problem, and I have started noticing it more clearly in myself. The brain is very good at conserving effort. Once there is a tool that can finish the sentence, organize the thought or resolve the ambiguity, there is very little friction preventing you from handing that work over earlier and earlier.

After a while, you begin to notice small changes. Words that once came naturally can take longer to arrive. You might open a prompt before you have really tried to work out what you want to say, or reach for assistance at the first sign of uncertainty instead of staying with the problem for a while. That can be unsettling when the thing becoming less immediate is part of a craft you have spent years developing.

My sense, though, is that these abilities are more dormant than lost. When I deliberately sit down and write without assistance, or stay with a design problem long enough to work through the uncertainty myself, I can feel those faculties returning to use. There is some friction at first, but that seems less like evidence of something disappearing and more like what happens when a part of the mind has simply been asked to do less for a while.

The risk, for me, is not that the tool suddenly takes away a faculty. It is that convenience gradually changes my habits without me noticing, until I stop practicing things I still care about being able to do. That is a much more ordinary problem, and probably a more manageable one.

I am not trying to outsource thinking wholesale, and I do not think avoiding powerful tools somehow protects the integrity of the work. I am trying to become more deliberate about which parts of the process I hand over and which ones I continue to exercise myself. I care about those faculties enough to keep using them, and I suspect that the presence of these tools may actually make that practice more conscious than it was before.

There is a strange possibility in that. A tool can make a faculty easier to neglect while also making you more aware of the value of having it. You begin to notice what used to happen almost automatically: remembering, phrasing, staying with ambiguity, working through uncertainty, developing an instinct slowly enough that it becomes your own. Sometimes you only really see a capacity once you feel yourself reaching for it and finding that it is not as immediate as it used to be.

That does not make me want to retreat from the tools. It makes me want to be more intentional about what I keep alive alongside them. If some of these faculties can weaken through disuse, they can also be strengthened through deliberate use. Writing can become practice for writing again. Memory can be exercised because I decide not to retrieve everything instantly. A model can help me inspect my thinking without always doing the thinking for me.

Maybe reclaiming our faculties becomes part of learning how to live with these systems. Not because the old way was purer, or because difficulty is inherently virtuous, but because there are abilities I still want to possess even when I no longer strictly need them. The easier it becomes to hand something over, the more deliberate the choice to keep practicing it has to become.

The craft remains available, but access to it still depends on use. If I want to keep writing, thinking and designing in a way that feels like my own, then some of the work still has to remain mine.