What is vibe coding and why does it get so much hate?

The Engadget piece unpacks "vibe coding" — the practice of writing software by describing intent in natural language and letting AI generate the code — and argues that the term's current reputation as synonymous with lazy, reckless development is a distortion of its origins. What began as a specific workflow built on iterative feedback loops has been flattened into a pejorative, obscuring a genuinely new mode of human-computer interaction. There's a familiar pattern at play here: a new tooling paradigm arrives, gets a catchy name, and within weeks the term is weaponized by both sides of a culture war. Vibe coding is the latest victim. The original framing — popularized by Andrej Karpathy in early 2025 — wasn't "prompt and pray." It described a tight loop: describe what you want, let the model write code, run it, hit errors, paste them back, iterate. The "vibe" referred to operating at a higher level of abstraction, trusting the feedback cycle rather than reading every line. That's not laziness; that's delegation with verification. But the public conversation rarely operates at that level of nuance. Critics seized on the worst-case scenarios: production code written by people who can't read it, security holes from hallucinated APIs, a generation of "developers" who are really just prompt jockeys. Some of those criticisms are legitimate — vibe coding without review is genuinely dangerous in production environments. But the reaction has been so allergic that any use of AI-assisted coding now risks being dismissed as unserious, which is its own kind of cargo cult thinking. The deeper issue is that vibe coding exposes a long-simmering tension in software engineering: is the craft the code itself, or the outcome the code produces? For decades, the industry has valorized the former — the elegant solution, the clever abstraction. Vibe coding inverts that hierarchy. It says the outcome matters more than the mechanism, and that the ability to articulate intent is a form of engineering skill. That's threatening to people whose identity is tied to the mechanics, and it's why the backlash has been so visceral. What gets lost in the shouting is that vibe coding is a spectrum, not a binary. There's a world of difference between using AI to scaffold a prototype you fully understand and shipping a critical service you can't explain to anyone. The former is a productivity revolution; the latter is malpractice. The hate directed at vibe coding is really hate directed at the latter, but it's been misapplied to the former. The term deserves rehabilitation — not because it's above criticism, but because the criticism should be aimed at the misuse, not the method. The most interesting question the article gestures toward is what happens to the next generation of developers. If vibe coding becomes the norm, do we lose the deep understanding that comes from writing code by hand? Or do we free up cognitive bandwidth for higher-level problems — architecture, user experience, business logic? History suggests it's not either/or. Every generation of tooling has been accused of dumbing down the craft, from compilers to high-level languages to frameworks. Each time, the industry adapted, and the bar for what counts as "real" work moved. Vibe coding is just the latest line in that long, contested sand.