Nvidia CEO: "AI Fluency is the New Coding"

- At the World Government Summit in Dubai, Jensen Huang argued that the rise of AI means programming is no longer a vital skill and that human language will become the new programming language. He suggested that time is better spent developing expertise in other fields like biology, education, or manufacturing. - This position counters the long-held advice from tech leaders over the past 10-15 years who have emphasized the importance of learning to code. Huang's argument is that it is the industry's job to create technology where no programming is necessary. - The AI-powered coding tool being used internally at Nvidia is a customized version of Cursor, an integrated developer environment from Anysphere Inc. that focuses on AI-assisted code design. While Nvidia claims a threefold increase in code output, the bug rate has reportedly remained flat. - Industry analysts are cautious about using "lines of code" as a primary measure of productivity, emphasizing that true software quality is better measured by stability, maintainability, and impact on end-user performance. - John Carmack, co-founder of id Software, supported Huang's view, stating that "problem-solving is the core skill" and that the discipline required by traditional programming will remain valuable but won't be a barrier to entry. - The shift toward AI-assisted development is expected to create new roles and require upskilling for a majority of software engineers. Gartner predicts that by 2027, 80% of engineers will need to adapt their skills for AI integration. - At OpenAI, 95% of engineers use the AI coding tool Codex, and those who have embraced these tools open 70% more pull requests, with the average code review time dropping from 10-15 minutes to 2-3 minutes. - Despite the productivity claims, some research indicates a potential "perception gap." One study found that experienced developers using AI tools took 19% longer to complete tasks, even though they believed they were working faster.