The Masochism of the Manual Syntax
There is a specific type of person who looks at a tool designed to make their life easier and says, "No thank you, I prefer the pain." We see it in the guy who buys a fixed-gear bicycle to commute in a city built entirely on 45-degree inclines. We see it in the people who shoot 35mm film and then wait three weeks to see a blurry photo of their own thumb. And now, we are seeing it in the hobbyist programming community, where using an LLM to write a function is increasingly viewed as the digital equivalent of using a cheat code in a game of Solitaire.
These are the 'Artisanal Coders.' They treat Python like it’s a temperamental sourdough starter. If they didn't hand-type every single character of a boilerplate class definition, did they even really code? To them, an AI-generated suggestion is a personal insult, a shortcut that robs them of the holy sacrament of banging their head against a mahogany desk until they realize they forgot to import a library. They don't want a solution; they want a struggle.
I recently saw a forum post where a user argued that using GitHub Copilot is "spiritually hollow." Spiritually hollow! We are talking about JavaScript, Brenda. It’s a language held together by duct tape and prayers. There is no soul in a for loop, and yet, here we are, treating a linter like it’s a desecration of a sacred text. The friction isn't a bug anymore; it’s a feature. It's the "analog warmth" of a segmentation fault.
The Mechanical Watch of the Digital Age
Think about the mechanical watch. It is objectively worse at telling time than a $10 Casio or the phone glued to your palm. It loses three seconds a day, it stops if you don't move your wrist enough, and it costs as much as a used Honda Civic. But people love them because they are 'honest' machines. Hobbyist coders are now treating a manual C build process with the same reverence. They want to see the gears turning, even if those gears are grinding their sanity into a fine powder.
Photo by DΛVΞ GΛRCIΛ on Pexels
This movement is a direct reaction to the 'Push-Button' era. When you can ask a chatbot to "write a weather app in React" and it spits out 400 lines of functional code in ten seconds, the magic dies for the hobbyist. If anyone can do it, then doing it isn't a personality trait anymore. To reclaim their identity, they have to move the goalposts into a swamp. They are intentionally choosing the most difficult path because the difficulty is the only thing that proves they were there.
It’s the "Slow Coding" movement. It’s about the journey, not the destination. Specifically, it’s about a journey where you get a flat tire every four miles and refuse to use a jack because you want to "understand the weight of the car." I’ve seen developers brag about not using autocomplete. Autocomplete! That’s like a carpenter bragging that he refuses to use a measuring tape because he prefers to judge the length of a board by the 'vibes' of the wood.
Debugging as a Luxury Spa Treatment
In the professional world, a bug is a catastrophe that costs $50,000 an hour in lost revenue. In the hobbyist world, a bug is an afternoon activity. It’s a crossword puzzle that hates you. By rejecting AI, these developers are turning the act of debugging into a high-end sensory experience. They talk about "tracing the logic" the way wine snobs talk about the "tannins" in a 1982 Bordeaux. "I really enjoyed the way this memory leak forced me to re-evaluate my entire life structure," said no one ever, except these guys.
- Phase 1: Denial. "I don't need an LLM, I have the documentation open in another tab."
- Phase 2: Anger. The documentation hasn't been updated since 2014.
- Phase 3: Bargaining. "If I just change this
intto along, the universe will stop screaming at me." - Phase 4: Enlightenment. Six hours later, the code runs. The dopamine hit is equivalent to winning the lottery, despite the fact that all the code does is make a digital cat meow.
There is a perverse pride in knowing that you suffered for your craft. If an AI writes the code, the dopamine hit is gone. You’re just a manager. And nobody goes into hobbyist programming because they want to feel like a middle manager at a mid-sized insurance firm. They want to feel like Neo in The Matrix, even if they’re actually just a guy in sweatpants trying to figure out why his CSS won't center a div.
What This Actually Means
We are witnessing the birth of "Heritage Computing." Just as people started baking bread during the pandemic because they missed the tactile reality of physical work, developers are retreating into manual syntax to escape the feeling of being a mere prompter for a machine. It’s a search for agency in a world where the machines are increasingly doing the thinking for us. Even if that agency involves spending three days trying to link a library that an AI could have fixed in a heartbeat.
Ultimately, this isn't about efficiency. It’s about the ego. We want to know that if the power goes out and the LLM servers melt into a puddle of silicon, we are the ones who can still build the fire. It’s silly, it’s inefficient, and it’s deeply human. We would rather be frustrated and in control than productive and redundant.
So, let them have their artisanal, hand-crafted, locally sourced syntax errors. Let them compile their kernels by hand and write their assembly on stone tablets. Just don't expect them to finish their projects any time this decade. They’re too busy enjoying the 'texture' of a stack overflow error to actually ship anything.
Quick Answers
Is using an LLM actually 'cheating'?
Only if you consider using a washing machine 'cheating' at doing laundry; you're still getting the job done, you just don't smell like a river anymore.
Why do hobbyists hate Copilot so much?
Because it turns a mysterious, wizard-like craft into a grocery list, and nobody wants to admit their 'magic' is just a very fast autocomplete.
Should I stop using AI to write my code?
Only if you have an abundance of free time, a high tolerance for migraines, and a desperate need to feel superior to strangers on Reddit.
