The man who coined the term 'vibe coding' says code written by AI can still be 'awkward' and 'gross' (www.businessinsider.com)

Andrej Karpathy, former Tesla AI chief and co-founder of OpenAI, recently expressed concerns about the quality of AI-generated code during a talk at Sequoia Capital. He described such code as "bloaty," "brittle," and often poorly structured, emphasizing that it still requires significant human oversight to ensure quality. Although Karpathy popularized the term "vibe coding," which refers to a development style heavily reliant on AI, he noted that current AI capabilities still leave much to be desired in terms of code aesthetics and functionality. He highlighted that while AI can assist with coding, it isn't yet capable of independently producing clean, efficient code. This statement comes amid growing unease in the tech community over the security vulnerabilities associated with AI-generated code. Recent incidents, such as those involving the Swedish platform Lovable, highlighted potential risks, including unintended data exposure. As vibe coding gains traction and influences hiring and employee evaluation practices, experts warn against overreliance on AI for coding tasks due to the persistent issues of code quality and reliability. Karpathy's insights reflect a broader discourse on the future of AI in software development, suggesting that while AI can significantly assist programmers, human expertise remains crucial for high-level decision-making and code integrity.