John Borthwick on AI Agents, the Death of Apps, and the Future of Software

John Borthwick argues that AI is making the founder more important, not less, because the cost of building technology is falling toward zero and the job is shifting from coding to steering. He says the best founders now need emotional attachment, adaptability, and the ability to survive repeated setbacks, which is why he probes them with questions like why they named their company what they did. He also says betaworks has always been more product-first than pattern-matching-driven, preferring close proximity to real products and real users over abstract founder heuristics.

He explains why betaworks chose a venture studio model and later added a fund structure: it wanted to do “power entrepreneurship” in related theses, run small thematic accelerator camps, and still be able to write pre-seed checks when a good company appears outside a camp cycle. Borthwick says this structure helped betaworks get early conviction in areas like conversational interfaces and AI agents, including Hugging Face and Granola. In his telling, the Hugging Face team was unusually thoughtful about narrow first use cases, then moved deeper into models as the stack changed.

Borthwick’s broader view of AI is optimistic but not naïve. He believes open and closed models both need to exist for a resilient ecosystem, and he is more worried about near-term security failures, sandbox breaches, and infrastructure weak points than about an abstract existential takeover. He says AI was “discovered” rather than invented, compares the current moment to Y2K-style plumbing work, and thinks white-collar workers should become deep users of the technology to stay antifragile.