Engineers are becoming artists, and artists are becoming engineers, at least according to Google AI Studio and Gemini API product lead Logan Kilpatrick. This is all thanks to AI. In a post on X that has racked up more than 627,000 views, Kilpatrick captured what he sees as a fundamental shift in how creative and technical work now overlap. “Thanks to AI, engineers have become artists,” he writes, “and all artists are becoming engineers, in a good way.”The comments reflect a growing sentiment in Silicon Valley that AI coding tools are erasing traditional boundaries between disciplines. Google CEO Sundar Pichai recently said that AI now writes more than 30% of new code at Google, and Anthropic CEO Dario Amodei claimed that Claude generates 90% of the company’s code. Nvidia CEO Jensen Huang went further, saying he wants engineers to stop writing code altogether and instead focus on “discovering undiscovered problems.” At a recent all-hands meeting at Nvidia, Huang reportedly pushed back against a manager who told his team to cut back on AI usage, asking bluntly, “Are you insane?”
AI tools expand access beyond traditional programmers
Kilpatrick’s observation echoes Pichai’s recent comments that “vibe coding,” the use of AI to build software with minimal technical knowledge, is making development “exciting again” and accessible to non-technical employees. People from various industries, from human resources professionals to accountants, are now using tools like ChatGPT, Gemini, Claude, and Cursor to prototype apps and automate workflows without formal coding training.
Reality check: Productivity gains remain uncertain
But optimism is not universal. Michael Truell, CEO of Cursor, the $29 billion AI coding assistant that Nvidia has hired across the company, warned developers not to blindly trust AI-generated code. “If you close your eyes and let AI build something on unstable foundations, things start to fall apart,” he told Fortune magazine. This skepticism is supported by research. The METR study found that AI assistants actually reduced the productivity of experienced developers by 19%, even though participants expected significant improvements.Even Andrei Karpathy, the former Tesla AI director who coined the term “vibecoding,” admitted that his recent project was “basically all handwritten” because the AI agent “didn’t work well enough.” He warned his fellow programmers: “I’ve never felt so far behind. This profession is being dramatically refactored.”
