An old man in Calcutta would walk to get water from a well every day. He’d carry a clay pot and lower it by hand slowly, all the way down, careful not to let it hit the sides of the well and break.

Once it was full, he’d raise the pot slowly and carefully again. It was a focused, time-consuming act.

One day, a traveler noticed the old man engaged in this difficult task. More experienced with mechanics, he showed the old man how to use a pulley system.

“This will allow the pot to go straight down quickly,” the traveler explained, “then fill with water and come back up, without hitting the sides. It’s much easier and the pot will be just as full with much less work.”

The old man looked at him and said, “I think I’m going to keep doing it the way I always have. I really have to think about each movement and there’s a great deal of care that goes into doing it right. I’d imagine if I were to use the pulley, it would become easy and I might even start thinking about something else while doing it. If I put so little care and time into it, what might the water taste like? It couldn’t possibly taste as good.”

— The Creative Act (pp. 93-94), by Rick Rubin

The reason I was initially drawn to programming is because it is a craft. Sure, it might sound goofy to compare code to traditional crafts, but programming shares the same core elements: building, problem solving, and creation. There was something truly satisfying about writing and understanding code, searching for arcane information in the depths of Stack Overflow, and seeing the fruits of many hours of labor in the form of a website. But now we all reach for the pulley. (Myself included.) It is a temptation that is near-impossible to ignore, and for economic reasons—that is, to remain competitive—it has become a necessity rather than a choice. What gets lost in this efficiency? The old man in the story communicated that the process (even of something mundane, like fetching water) was inseparable from the quality of the outcome. When I used to debug a particularly stubborn piece of code, I'd spend hours tracing through logic, reading documentation, testing theories. It was frustrating, sure, but that frustration taught me how the system actually worked. I built mental models through trial and error. Now, when AI solves the problem instantly, I get the solution but lose the deep understanding that comes from wrestling with it myself. ## We are becoming cognitively bankrupt This is not just limited to programming. In this new era of AI, we are bombarded with pulley systems for every facet of life: asking Grok to evaluate reliability of information, asking ChatGPT to write our emails, and asking Cursor to write our code. Each of these conveniences fundamentally changes how we engage with the world. When we ask AI to evaluate information for us, we gradually lose our ability to discern bias, check sources, or think critically about what we read. The muscle of skepticism atrophies. When we let AI write our emails, we stop learning how to communicate nuance, handle difficult conversations, or develop our own voice. The careful work of choosing the right words—words that actually represent what we mean—gets outsourced. It's the reason the most powerful tools at our disposal are making us [cognitively bankrupt](https://www.media.mit.edu/projects/your-brain-on-chatgpt/overview/). Consider what happens when we stop doing the small, seemingly mundane tasks that actually build our capabilities. Reading dense technical documentation teaches us patience and thoroughness. Debugging teaches us systematic thinking. Writing difficult emails teaches us empathy and precision. Each time we reach for the AI pulley, we're not just saving time—we're choosing not to develop these deeper capacities. ## I swear I’m not a luddite I am by no means a luddite, and think most criticisms of AI are not well-founded. But I am reflecting often on how I feel about the act of “writing” code (which is likely now a misnomer for most programmers). I'm sure many people—especially artists who have invested decades in mastering their craft—share this sense of confusion, frustration, or unease. The unease likely stems from many sources: the obvious fear of losing jobs, but also the deeper fear of losing identity. When a painter can no longer distinguish their work from AI-generated art, when a programmer spends more time prompt engineering than problem solving, when a writer becomes primarily an editor of AI output—what are we? The craft that defined us becomes something we manage rather than something we do. ## Pulleys go both ways The paradox here is that these new pulley systems can make craftsmen feel purposeless, but they can make the same craftsmen feel liberated with new opportunity. On my portfolio site, I wanted to create a consistent series of icons in a 3D style (which you can see below). Two years ago, this would require a tremendous learning effort involving multiple tools (or paying someone to do it for me). Today, it was simple as one prompt in Recraft. *(There is no real point in this section, I am simply dwelling on if this makes me a hypocrite.)* From an economic perspective, those that succeed will be those who bias toward action and are [loud](/writing/audacity-goes-the-furthest). Carefully-attained knowledge is not the moat it once was—attention, distribution, and utilization of tools have replaced it. ## Will "humanmade" be the new "handmade"? How essential is the act of creation to our collective well-being? Probably not too much anymore—creation and craft are declining in relative importance just as we enter a hyperconsumption era. Few care about craft anymore, and even though I would argue this makes them feel worse about their lives, most are not of this belief (or have not dwelled on it enough for it to inspire action). We're witnessing something unprecedented: the potential end of creation as a widespread human activity. For most of human history, people made things—clothes, food, tools, shelter, stories, music. Even in the industrial age, most people still had hobbies that involved creating something with their hands. Now we're moving toward a world where creation becomes the domain of machines, and humans become primarily consumers and curators. This shift has psychological implications we're only beginning to understand. I believe that for most people, creating things—even simple things—contributes to a sense of purpose and life satisfaction. The act of making something that didn't exist before, of solving problems with your hands and mind, seems to fulfill some deep human need. When we outsource all creation to AI, what happens to that part of us? There is a sad, dystopian, yet plausible future where “humanmade” becomes the new “handmade”—software, art, media that is created exclusively or predominantly by humans. But these will be relics precisely because they are novel. Perhaps we’ll find new ways to engage in “craftsmanship”—prompts and workflows and orchestrations of existing tools. And perhaps a version of me would have written this same blog post 200 years ago, when we were on the verge of major disruption of traditional craftsmanship by factory machines—lamenting that 'handmade' was dying while mass production took over.

Writing

Personal

AI is the Pulley

Another Reason the Internet Sucks

Audacity Goes the Furthest

No Breaks While Running

The Marble and Its Sculptor

Turbulence is Expected

Actors and Extras

Technical

Integrated Security

Making Earth Habitable