AI Makes Experienced Developers Nineteen Percent Slower

The emergence and consequences of vibe coding, a practice where developers use AI agents to generate software via natural language prompts rather than manual programming. While Andrej Karpathy originally coined the term for low-stakes hobby projects, research from METR suggests that technical workers now report significant productivity value from these tools despite high rates of self-reported errors. However, this shift has led to a "vibe coding hangover" characterized by a massive accumulation of technical debt, severe security vulnerabilities, and a decrease in code maintainability. Industry experts warn that the speed of AI-driven development often outpaces human review, potentially compromising the long-term health of open-source ecosystems. Organizations are now being urged to implement formal governance and accountability structures to mitigate the risks associated with blind reliance on automated output. Ultimately, the texts describe a transition where the developer’s role is evolving from a writer of code to a high-level orchestrator and validator of AI-generated systems.