We asked over 150 software engineers about vibe-coding. Here's what they said.

The emergence of AI technology has transformed the landscape of coding, prompting a wave of discussions among software engineers about productivity and job security. The term "vibe-coding," coined by Andrej Karpathy, refers to the practice of leveraging AI for coding tasks. This concept has gained significant traction and was recognized as the Collins Dictionary's Word of the Year for 2025. In a recent survey of 167 software engineers, opinions varied on how vibe-coding has impacted their work. Approximately 46.9% of respondents—75 engineers—indicated they felt they were keeping pace with the rapid changes in the industry. Meanwhile, 30 engineers reported feeling ahead of the curve, whereas 27 expressed feeling left behind. Interestingly, 17.5% of respondents—28 engineers—opted out of AI coding tools entirely, citing concerns about their effectiveness and a steep learning curve. Despite the survey's non-scientific nature, it sheds light on the diverse perceptions of software engineers regarding AI tools. Follow-up discussions with eight engineers revealed a spectrum of experiences with AI code editors. Most found them beneficial, with applications varying from occasional assistance to essential tools for their workflow. Ryan Shah, a recent computer information technology graduate, expressed that while he sometimes questions his coding skills, the ability to read code remains crucial in a rapidly evolving environment. Conversely, Javanie Campbell, CEO of DevDaysAtWork, cautioned against over-reliance on these tools, warning that those who treat AI as the ultimate authority might face job insecurity. The debate within the engineering community centers on the potential impact of AI on job opportunities: some believe it may lead to workforce reductions, while others view it as a supplement to human expertise. Ryan Clinton, a software developer, initially feared that vibe-coding would jeopardize his career but has since come to terms with its integration into his work. He noted that while AI can assist with coding tasks, human oversight is still essential. Barry Fruitman, an Android developer, believes the full effects of AI on the job market won't be felt for several years, expressing hope that current fears are exaggerated. On the productivity front, Ed Gaile reported significant gains in his workflow, stating that AI tools have drastically reduced context-switching. However, a July METR study revealed a paradox: while AI-assisted developers completed tasks more slowly, they also spent more time on code review, leading to questions about overall productivity. Shawn Gay, an R&D manager, voiced his concerns about adapting to the rapid changes in software development, feeling the pressure of keeping his skills relevant. Similarly, Gus De Souza acknowledged that while AI can speed up coding, it often requires more time for reviewing outputs. Vibe-coding, as defined by Karpathy, encourages developers to fully embrace AI's capabilities while keeping foundational coding skills intact. However, not all engineers identify with this approach. Lara Fraser, a data analyst, expressed skepticism about relying solely on AI tools, emphasizing that coding is not just about creation but also about maintenance and problem-solving. As the conversation around vibe-coding continues, it is clear that the integration of AI in software development presents both opportunities and challenges, urging engineers to strike a balance between innovation and foundational skills.