John Carmack recently tweeted a warning that felt like a glitch in the simulation: stop trying to be the 'out of touch Kung Fu master' who spends decades perfecting a single strike. For a guy who famously spent the 90s optimizing assembly code for Quake and Doom to squeeze every millisecond out of primitive hardware, this is a massive pivot. He’s essentially telling a generation of developers that their pride in deep, machine-level mastery is becoming a liability rather than an asset. It makes me wonder if we’ve reached the point where the 'how' of technology has finally been completely swallowed by the 'what.'
The Loneliness of the Perfect Loop
There used to be a specific kind of romance in writing the perfect, elegant algorithm. It was about a private relationship between a human mind and the silicon, a struggle to make the hardware do something it wasn't designed to do. Carmack was the patron saint of this struggle. When he speaks about the 'Prestige Trap,' he’s talking about the dopamine hit developers get from solving a hard technical problem, even if that problem doesn't actually move the needle for the person using the software.
I find myself wondering if this 'prestige' was actually the only thing keeping the job human. If you take away the craft—the clever workarounds, the manual memory management, the fight with the compiler—what are you left with? You’re left with a translation layer between a business requirement and an API. It’s efficient, sure. It’s outcome-oriented. But is it still a vocation, or is it just data entry with higher stakes? We’re moving into an era where the machine is so powerful and the abstractions are so thick that 'mastery' looks more like knowing which AI prompt to use than knowing how a cache miss affects performance.
The $100 Billion Abstraction Layer
Consider the sheer amount of capital currently being incinerated to make sure no one ever has to write 'hard' code again. We are spending billions to build LLMs that can hallucinate a functional React component in three seconds. In this environment, the 'Kung Fu master' who wants to spend a week hand-tuning a database schema looks less like a genius and more like a hobbyist who doesn't realize the factory has already moved to another country.
- The cost of 'good enough' software has plummeted to near zero.
- The value of 'perfectly optimized' software is increasingly restricted to a tiny handful of high-frequency trading firms and aerospace companies.
- We are prioritizing speed of iteration over the longevity of the code base.
It’s a strange trade-off. We gain incredible speed and the ability to solve massive problems without needing a PhD in computer science. But I worry about what happens when the people who understood the underlying systems are gone. If everyone is a 'product-market fit' specialist using AI-generated blocks, who fixes the AI when it breaks the foundation? We are building a skyscraper and deciding that the people who know how to mix concrete are 'out of touch.'
The Identity Crisis of the 10x Developer
If you spent twenty years defining yourself by your ability to solve the unsolvable, Carmack’s advice feels like a personal attack. The 'Kung Fu' metaphor is sharp because it implies that these technical skills are now just performative—a beautiful dance that doesn't actually win the fight. The new fight is simply 'does this product exist and does it work?' This shift forces a generational identity crisis because it devalues the very thing that made the tech industry a meritocracy in the first place: raw, undeniable technical competence.
I’m curious about what the 'new' prestige looks like. If it’s not the code, is it the vision? Is it the ability to stitch together disparate AI agents into a coherent system? We’re seeing a transition from 'builders' to 'conductors.' A conductor doesn't need to know how to play the oboe at a world-class level, but they need to know what the oboe is supposed to sound like. But a conductor who has never played an instrument is a very different kind of leader than one who has.
What This Actually Means
We are witnessing the final decoupling of 'software engineering' from 'computer science.' For a long time, we treated them as the same thing, but Carmack is pointing out that they are diverging. Engineering is now about outcomes, delivery, and utility. Computer science is becoming a niche academic pursuit for the people who build the tools the engineers use. If you’re a developer today, you have to decide which side of that line you want to live on, and the 'engineering' side is getting very crowded with people who don't care about the beauty of the code.
This isn't necessarily a bad thing for the world. More problems will get solved faster. More people will have access to the power of creation. But for those of us who grew up admiring the 'masters,' there’s a genuine sense of grief. We are trading the depth of the individual for the breadth of the system. We’re trading the artisan for the architect.
The 'Kung Fu' era of tech was defined by the individual hero—the lone coder who stayed up all night to save the project with a brilliant hack. The next era will be defined by the orchestrator who manages a swarm of automated processes. It’s more effective, but I can’t help but feel it’s a lot less cool. The master has been told to put down the sword and pick up a clipboard.
Quick Answers
Is John Carmack saying technical skills don't matter anymore?
No, he’s saying that technical mastery shouldn't be an end in itself if it doesn't serve the product. He’s warning against 'playing' at engineering instead of solving actual problems.
What is the 'Prestige Trap'?
It’s the tendency for experts to focus on complex, high-status technical challenges that earn them respect from peers but offer diminishing returns for the actual users of the software.
Should new developers still learn the basics of how computers work?
Yes, but with the understanding that those skills are now 'emergency' skills rather than daily ones. You need to know how the engine works so you can fix it, not so you can manually pump the fuel while driving.
