This is quite interesting.
You all know Karpathy, right? The AI heavyweight who’s worked at both Tesla and OpenAI. He recently dropped a bombshell in the developer community: “After coding for 20 years, using AI now feels like cheating.”

Honestly, I laughed when I first read that. Isn’t this the exact dilemma we face every day? When using Copilot, we’re simultaneously thrilled (“Holy crap, it can auto-complete this?!”) and guilty (“Who actually wrote this code?”).

1. From “Vibe Coding” to “Feels Like Cheating”
Karpathy previously introduced the concept of “vibe coding,” which roughly means letting AI sense your programming intent and semi-automatically write code. At the time, everyone cheered, calling it the future. Now, he’s backtracking, saying the model has issues.

This is painfully relatable. When AI-assisted programming first emerged, everyone was hyped. But once the novelty wore off, the real problem surfaced: Efficiency improved, but the sense of accomplishment from coding vanished. It’s like using cheats in a game—you finish faster, but it’s utterly unsatisfying.

2. That 110-Task Experiment
The article mentions a wild experiment—an AI Agent independently completed 110 coding tasks, including hardcore stuff like setting up a home surveillance system. Sounds impressive, right? But anyone in tech knows: Getting it to run and making it production-ready are two different things.

I’ve seen too much AI-generated code:

  • It runs, but it’s spaghetti code that’s a nightmare to maintain.
  • Zero comments—three months later, even its creator wouldn’t recognize it.
  • Crashes at the slightest edge case.

So here’s the reality: AI can handle 80% of the boilerplate code, but humans still have to clean up the remaining 20% of devilish details.

3. What Are We Really Afraid Of?
Saying it “feels like cheating” reveals programmers’ deeper anxieties:

  • Fear of becoming “AI code proofreaders.”
  • Fear of losing technical judgment (just blindly following AI’s suggestions).
  • And the biggest fear—what if bosses think, “If AI can write code, why do we need you?”

But here’s a different perspective. When cars replaced horse-drawn carriages, carriage drivers became drivers. Now that AI handles repetitive tasks, shouldn’t we focus on cooler things? Like designing architectures, optimizing algorithms, or solving complex problems AI can’t handle?

One Final Observation
The Chinese community’s take on this discussion is uniquely pragmatic. While Western debates revolve around “Will AI replace programmers?”, we’re more concerned with “How to use AI to stay competitive.” Maybe that’s the difference between practitioners and theorists?

In the end, Karpathy’s statement is a wake-up call for everyone: The more powerful the tool, the more important it is to know when not to use it. Just like you wouldn’t forget that 1+1 equals 2 just because you have a calculator, right?

(The End)