The landscape of software development is undergoing a rapid transformation, with generative artificial intelligence increasingly woven into the fabric of daily coding tasks. Routines that once consumed days or even weeks for developers can now be executed in moments, leading many to embrace what some term “vibe coding” over the meticulous, line-by-line approach of traditional programming. However, this shift is not without its vocal critics, among them Markus Persson, the billionaire creator of Minecraft, widely known as “Notch.” Persson recently voiced a stark warning, asserting that anyone advocating for the use of AI in writing code is either “incompetent or evil.”
Persson’s concern, articulated in a recent social media post, extends beyond a simple resistance to technological advancement; it delves into fundamental anxieties regarding autonomy, creative control, and the very essence of what defines coding as a discipline. He drew a parallel between AI-generated code and AI-written laws, arguing that both are fundamentally about logic, not merely the act of typing. The 46-year-old developer also suggested a deeper, more manipulative undercurrent to the push for AI integration, stating, “If you seriously believe that the same people that manipulate people via mainstream media (for clearly evil purposes) wouldn’t use subtle ways to use AI to manipulate us, you’re dumb. It’s being forced down our throats for a reason.” His unequivocal bottom line: “Don’t rely on AI for programming.”
This cautionary stance from Persson arrives at a pivotal moment, as a new generation of engineers enters a tech industry increasingly focused on headcount reduction. For these emerging professionals, prompting and “vibe coding” are not just fleeting trends but are rapidly being presented as viable replacements for established programming practices. The adoption rates of AI in major tech companies underscore this trend. Google CEO Sundar Pichai reported last April that over 30% of Google’s code is now AI-generated, a figure echoed by Microsoft CEO Satya Nadella, who stated his company’s AI-generated code falls between 20% and 30%. Leaders at AI-centric startups like OpenAI and Anthropic have even claimed their technology now produces 100% of their code.
Persson, who taught himself to code in Sweden before launching Minecraft in 2011—a game that has since sold over 300 million copies and was acquired by Microsoft for $2.5 billion—remains deeply skeptical. Despite Minecraft’s role as an entry point for many young people into programming and, more recently, AI experimentation, he advises aspiring tech developers to “Learn. Your. Craft. You are not safe to program if you don’t know how to do it,” likening reliance on AI to knowing only how to fly a plane without understanding how to land or take off.
While Persson stands firm in his conviction that foundational programming skills are indispensable, the tech industry itself is divided on what this foundation should entail in the AI era. GitHub CEO Thomas Dohmke, for instance, advocates for universal coding literacy, suggesting that every child should learn to code in school alongside subjects like physics and geography. This perspective underscores a belief that programming should be a core skill for everyone, not just future software engineers.
Conversely, figures like Andrew Ng, co-founder of Coursera and former head of Google Brain, openly challenge Persson’s views. Ng actively encourages developers to leverage AI as a coding assistant, advising against “the old way” of coding by hand. He believes that integrating AI into the coding process will significantly enhance productivity and enjoyment across various job functions. Nvidia CEO Jensen Huang has taken an even broader view, asserting that AI will affect every job “immediately,” and that individuals risk being outcompeted by those who effectively utilize AI, rather than by the AI itself. This ongoing debate highlights the profound philosophical and practical questions facing the tech world as generative AI continues to reshape the very definition of coding.
