'AI Coding Is Gambling': Why the Infinite Code Machine Leaves Developers Feeling Empty
A developer who has been coding with AI since November 2025 — when "we all noticed it got really good" — offers a candid reflection on what we've gained and, more importantly, what we've lost.
The Gambling Proposition
The author identifies a core dynamic: getting into a state where any change to your entire codebase is trivial to make is intoxicating. Previously, developers weighed the cost of each task — research, reading forgotten code, reconnecting to old thinking. Now, the AI can handle it. Or at least pretend to.
"But this doesn't really resemble coding. An act that requires a lot of thinking and writing long detailed code. Both parts are technically here, but the first isn't essential and the second can be minimal."
Why It Maps Onto Gambling
The argument is that AI coding perfectly mirrors the mechanics of a slot machine:
- Variable reward — sometimes it works perfectly, sometimes it's "surprisingly wrong"
- Low-friction iteration — just ask again, pull the lever once more
- The appearance of productivity — you're "doing something" even when the results are shallow
- Compulsive engagement — the author notes they left Claude Code running for 8 days and accumulated "an 8+ year portfolio of projects"
What's Missing
The author divides tasks into "good for the soul" and "bad for it." Coding traditionally fell into the first category — figuring things out, finding clever fixes, making things work. With AI, that process is outsourced:
"My job went from connecting these two things being the hard and reward part, to just mopping up how poorly they've been connected. It's deeply unsatisfying."
The part that's "best for the soul" — figuring out how something works for you, finding the clever fix, getting it working — is now handled by the machine. What remains is cleanup and verification.
Not Anti-AI
Importantly, the author isn't calling for AI's abolition or warning about job losses. They acknowledge AI coding genuinely requires skill ("you really gotta know what you're doing"). The point is more subtle: the experience of coding has fundamentally changed, and something valuable was lost in the transaction.
The Bigger Picture
This piece resonates with a growing conversation about the qualitative experience of working with AI tools. The metrics say we're more productive — more code shipped, more projects completed. But the subjective experience tells a different story about satisfaction, craftsmanship, and the joy of problem-solving.
Source: VS Notes | HN Discussion