Will Knight, a senior AI correspondent for Wired, successfully built a recursive AI system—self-improving technology that OpenAI, Google DeepMind, and Anthropic have invested billions in—without frontier lab resources, signaling a shift in who controls AI development. Knight's experiments prove that self-improving AI no longer requires massive compute clusters and research teams affordable only by those companies, as the recursive improvement techniques are now replicable with accessible tools and modest resources. This development comes as OpenAI closed its latest funding round at a $150 billion valuation and Anthropic raised another mega-round, challenging the assumption that only frontier labs can build advanced AI. Knight built working systems exhibiting self-improvement capabilities, demonstrating that the democratization of AI-building-AI tools has crossed a critical threshold. For startups, the narrative that hundreds of millions in funding and a team of PhD researchers are needed to compete in AI is now obsolete, with expectations of specialized self-improving systems for vertical markets. Enterprise buyers should note that vendor lock-in from only three companies delivering frontier AI is now negotiable, as proprietary self-improving models trained on their own data become possible.