December 9, 2025
Vibes vs Bugs: pick a side
If You're Going to Vibe Code, Why Not Do It in C?
Internet melts down over AI vibe coding—half amazed, half screaming slop
TLDR: The author says AI “vibe coding” can build real apps and cheekily suggests doing it in C or even assembly. The comments explode: most say not, warn about unvetted “slop,” and propose safer guardrails like Rust, turning a simple question into a full-on culture war over how we write software.
An old-school professor confessed a guilty secret: “vibe coding” with AI actually works. He adores hand-written code—the kind that feels like solving a magical crossword—yet admits AI can ship real, complex apps. Then he poked the bear: “If you’re going to vibe code, why not do it in C—or even assembly?” Cue flamewar.
The top mood? Hard NO. Auntienomen slammed the premise with, “If the headline is a question, the answer is ‘No’.” Esafak warned that pure vibe coding is a joke, arguing they need to review and extend their own apps, not trust unvetted machine output. The business angle got spicy too: “Do you really want to sell code no human has checked?” Ouch.
Others tried to steer the chaos. Stared linked a debate—“Why AI Needs Hard Rules, Not Vibe Checks”—and pitched Rust, a “safety-first” language, as the vibe coder’s seatbelt. The anti-AI crowd cranked the sirens: bigstrat2003 called vibe output “a hot mess,” while hamzaawan said once AI slides into “slop,” you’re stuck digging holes and spinning up “a new repo” for every tiny fix.
Jokes flew: “Assembly? My wrists hurt already.” Verdict: the community’s split between dazzled experimenters and skeptical grown‑ups—both loud, both caffeinated today online.
Key Points
- •The author questions why AI-driven “vibe coding” couldn’t be done in low-level languages like C or x86 assembly.
- •They describe personal discomfort with vibe coding despite a strong passion for programming and teaching.
- •The author claims vibe coding can produce robust, complex systems, based on their own experience using an AI model.
- •They recount a progression from skepticism to practical use, finding AI effective for solving programming tasks and generating programs.
- •The article cites SICP’s Abelson and Sussman to emphasize that programs should be written for people to read, underscoring human readability.