Forget Coding, Get A Hobby: Anthropic Co-Founder Shares Surprising Career Advice For The AI Era- 'It's A

The co-founder of one of the world's most important AI companies has a message for workers anxious about being replaced by machines: stop learning to code, and start developing a hobby.

Jack Clark, who co-founded Anthropic — the company behind the AI assistant Claude — told BBC Newsnight this week that creativity and curiosity will matter more in the AI era than any single technical skill. The advice carries particular weight coming from someone at the center of the industry disrupting traditional career paths. After all, Anthropic recently disclosed that Claude now generates roughly 80 percent of the code used within the startup itself.

"If you're a young person trying to figure out what to do, developing a hobby is actually one of the most valuable things," Clark argued. "It teaches you to be curious, to explore, to develop passions outside of what's immediately productive. And that's exactly the kind of thinking that AI can't replicate."

The statement landed like a thunderclap in an education system that has spent the last decade pushing every student toward STEM fields. Coding bootcamps proliferated. Computer science became the most popular major at top universities. "Learn to code" became both career advice and cultural shorthand for economic adaptation. Now, one of the people building the technology that makes much of that coding obsolete is saying the path was never as stable as it appeared.

The reality behind Clark's advice is borne out by the data. Anthropic's own economics research, published in early 2026, found that AI systems are most effective at replacing routine cognitive tasks — precisely the kind of entry-level coding, data analysis, and content generation that recent graduates are often hired to do. Meanwhile, tasks requiring interpersonal judgment, creative synthesis, and domain expertise that spans multiple fields remain stubbornly resistant to automation.

This does not mean technical skills are worthless. The same Anthropic research shows that workers who combine technical literacy with deep domain knowledge — understanding both the tools and the problems they're solving — remain highly productive. The threat is to the purely mechanical coder, the person whose job is to translate a specification into syntax. That job is being eaten from both ends: AI writes the code, and the people who understand what the code should do increasingly interact with AI directly rather than through a human intermediary.

Clark's hobby prescription is more nuanced than it sounds. He is not suggesting people abandon career preparation for leisure. Rather, he is arguing that the kind of nonlinear, exploratory thinking that hobbies develop — whether that is woodworking, rock climbing, chess, or creative writing — builds cognitive flexibility that rigid skill training does not. In a world where the specific technical skills you learn today may be automated next year, the ability to adapt, to see connections across domains, and to ask interesting questions becomes the most durable professional asset.

The advice also reflects a broader shift within the AI industry itself. At Anthropic and similar companies, the most valuable employees are not those who write the most code. They are researchers who can frame novel problems, ethicists who can anticipate societal impacts, and product leaders who can identify needs that no one has articulated yet. The technical implementation is increasingly commoditized; the human insight is what differentiates.

For workers, the implications are both liberating and uncomfortable. Liberating because it means the relentless credential-chasing of the past decade — the second master's degree, the sixth certification, the latest framework tutorial — may matter less than cultivating genuine intellectual curiosity. Uncomfortable because hobbies and curiosity are harder to optimize for than technical skills. You cannot reliably turn a woodworking hobby into a six-figure salary the way you could with a coding bootcamp in 2021.

The labor market is already adjusting. Companies that once required specific programming languages on job listings are increasingly testing for problem-solving ability and communication skills instead. Google, Microsoft, and Amazon have all modified hiring practices in 2026 to emphasize adaptability over specific technical proficiency. Even consulting firms and banks, traditionally bastions of rigid credentialism, are experimenting with hobby-based interview questions and portfolio reviews that value breadth over depth.

The educational establishment is slower to adapt. Computer science enrollments remain near record highs, though growth has flatlined for the first time since 2015. Some universities, including Stanford and MIT, have begun offering interdisciplinary programs that combine technical training with philosophy, design, or social science. But the vast majority of institutions continue to teach coding as though the market for entry-level programmers will remain robust indefinitely.

Clark's framing also raises uncomfortable questions about inequality. Hobbies require time, resources, and often parental support that many families cannot provide. A child from a wealthy suburb who develops a passion for robotics club or debate team has a structural advantage over one from an under-resourced school that barely offers art class. If curiosity and creative exploration become the primary currency of the labor market, the gap between well-resourced and under-resourced students could widen further, not shrink.

What This Means For You: If you are mid-career, the shift toward valuing adaptability over specific technical skills is good news. Your accumulated experience, judgment, and professional network are precisely the kinds of assets that AI cannot replicate. Invest in connecting dots across domains rather than drilling deeper into a single specialization. If you are earlier in your career, or advising someone who is, the playbook has changed. Technical literacy remains important, but it is increasingly a baseline expectation rather than a differentiator. The people who will thrive are those who combine technical competence with genuine curiosity, creative thinking, and the ability to communicate across boundaries. Developing a hobby is not frivolous advice — it may be the most practical career strategy available.