There’s no shortage of LinkedIn-core opinions on how we should restructure our personal and professional lives in the wake of widely available text, image, and video models. Most of the takes are boring, but one reliably gets under my skin: the claim that taste is now more valuable than craft. The folks saying this (most of whom I suspect fancy themselves very tasteful) posit that when a machine can spit out anything we can dream up, and the gap between thinking and doing becomes small enough, we’ll no longer value the skill of being able to execute on an idea as much as we value the discernment of knowing what’s worth making in the first place. And further, they want us to be excited that we’re no longer bottlenecked by our own (or others’) skills; we can finally design, paint, write, animate, code, create at the speed of thought!
Practically, some of that is undeniable. Sora and Claude and Nano Banana are here, they are cheaper and easier than the alternatives of Maya and Upwork and Photoshop, and they do lower the barriers to entry for previously specialized fields. I cannot and will not dispute any of that. At the same time, though, the new focus on taste feels a little like a retreat to high ground as the generative AI waters keep rising. A few years ago we laughed at 8-fingered hands and said LLMs could never replace human artists or engineers. Now we define more abstract value propositions (taste, judgement, creativity) and confidently declare that these are the unique traits we bring to the table. But ultimately I don’t buy that these are defensible skills—data-driven personalization already defines the media we consume, and I have to imagine it’s only a matter of time before AI can successfully separate hits from duds for every demographic niche and subculture.
But my real problem with this elevation of the tastemaker over the craftsman is not that it’s naive. In fact, I don’t really care about whether it’s right or wrong in a literal sense. The reason I bristle when I hear people talking this way is that I fundamentally don’t believe that thought-making and thing-making are independent in the first place, and I worry there’s a lot to lose if we pretend they are. If we start treating ideas as something to be produced or evaluated without taking part in their realization, we turn a network of thought and action into a single-direction pipeline. Accepting that ideas are simply inputs to be given to some Do Something Machine robs us of the conscious and unconscious thought we’d have experienced if we’d gone and Done Something on our own. It also short-circuits the creative process, compressing the range of what we’re able to make at all.
What I’m getting at here is a pretty basic fact, which is that cognition happens in more varied ways than the conscious act of assembling sentences or pictures in your mind. Sketching some rough drafts of a design you’re planning, taking notes or doodling while you listen to a lecture, rearranging sticky notes on a whiteboard, stacking blocks, walking through a room before filming in it, laying out materials before building with them, shuffling a deck of cards, writing and benchmarking your own code, retooling or deleting a sentence that doesn’t quite work; each of these engages a quiet, lurking part of the mind in the creative process and affects what you’re making in ways major or subtle. Touching, moving, and creating involve the body in the process of cognition, turning the invisible mills that grind up raw experience to process, synthesize and connect information in your mind. It may change what you end up creating, or it may just change your perspective on it; either way, you’ve got something different than what you started with. Whether you’re aware that it’s happening or not, this internal process is as worthy of the title of “thought” as some upfront, focused brainstorming or problem-solving process.
This is why I don’t want to relinquish the responsibility of actually making things, or to devalue process and craft and skill. I’m not under some delusion that we can make prettier pictures than the generative models can (I, for one, certainly cannot). But if we let AI do all the boring parts of our creative or productive work, I fear we’re trading away an entire category of human experience and expression. By separating ideation from execution as if they’re two independent components in a supply chain, we may give ourselves more time to think, but we’ll be greatly limited in the variety of thought that we have access to.
Thinking is not solely a precursor to action, and doing something is not a rote bodily function performed in response to a prompt. Doing is thinking. Put another way, doing your own work, with your own mind and body, is the practice of a certain form of unconscious thought. And participating in this process is a joy! To be surprised by an arriving sentence or image, to put a name to a face at last, to be struck with a notion from the blue, to feel understanding click into place, to suddenly see a system as a whole and in its parts, to find a thread that connects two seemingly unrelated ideas is to experience a sublime gratitude and humility. “Aha,” you’ll say, and smile.