A designer says AI coding feels like a slot machine
A viral blog post argues that AI-assisted coding has turned programming into compulsive gambling — pulling a lever and hoping the output works instead of actually understanding the code.
A blog post titled "AI Coding Is Gambling" hit the front page of Hacker News this week with 77 points and dozens of heated comments. The core argument: using AI to write code feels less like programming and more like pulling a slot machine lever — and it might be making us worse at our jobs.
The addiction loop
The author — a designer who also codes — describes a pattern that anyone using AI coding tools like Cursor, Claude Code, or GitHub Copilot will recognize:
"Getting yourself in a state where any change to your entire codebase is trivial to make is intoxicating."
The argument goes like this: when AI can make any code change feel effortless, you stop weighing whether a change is worth making. You just keep asking the AI to try things — regenerate, tweak, regenerate again — hoping the next output will be the one that works. That's the slot machine.
Traditional coding forced you to think before typing because every change cost effort. That friction was actually valuable — it made you understand your code deeply. With AI, the friction disappears, and so does the understanding.
"Pretending to handle it"
The post's sharpest observation: AI-generated code often looks correct but isn't. The output appears complete and professional, but contains subtle bugs that only someone who understands the underlying logic would catch. The problem is that the "understanding" part — traditionally the most rewarding aspect of programming — has been delegated to the AI.
As the author puts it: "Am I actually more efficient and smarter, or is it because I'm just gambling on what I want to see?"
What developers lost (and gained)
The post draws a clear line between two types of coding work:
"Good for the soul" — Figuring out solutions, finding clever fixes, understanding how pieces connect. This is what made programming satisfying and built real expertise.
"Bad for the soul" — Mopping up how poorly AI-generated pieces connect. Debugging code you didn't write and don't fully understand. The tedious cleanup that follows a gambling spree.
The author isn't anti-AI. They're raising a question that a growing number of developers are quietly asking: is the speed boost worth what we're losing?
Why this resonates right now
This post comes at an interesting moment. AI coding tools are reshaping how software gets built, with studies showing AI can speed up certain tasks dramatically. But there's a growing counter-narrative:
A CMU study of 806 projects found quality concerns with AI-assisted code. The concept of "comprehension debt" — the hidden cost of code you didn't write and can't fully understand — is gaining traction. And Django just flagged AI-generated contributions as a growing problem for open-source projects.
The takeaway for AI tool users
If you use AI to help with coding, writing, or any creative work, the gambling metaphor is worth sitting with. The question isn't "does AI make this faster?" — it clearly does. The real question is: are you still learning and understanding, or are you just pulling the lever?
A practical middle ground: use AI for the boring, repetitive parts (boilerplate code, formatting, syntax lookup), but force yourself to write the logic yourself. That way you keep the speed boost without losing the understanding that makes you actually good at your job.
Related Content — Get Started with Easy Claude Code | Free Learning Guides | More AI News
Stay updated on AI news
Simple explanations of the latest AI developments