The AI economy is not creating a new economic order but accelerating the oldest one - feudalism with computational monopoly replacing land ownership. The concentration of AI capability in a handful of corporations, combined with the historical pattern of technological revolutions being captured by existing power structures, suggests a feudal outcome unless open source AI provides a structural counterforce. The question is not whether AI creates or destroys jobs, but who owns the intelligence infrastructure that will mediate all economic activity.

Ilya Sutskever’s admission that scaling AI will continue to improve capabilities but leave something important missing points to a fundamental gap between intelligence and wisdom. This essay explores why more capability without better judgment may simply accelerate humanity’s existing failures, and why the real bottleneck in AI development isn’t technical but human.

AI development is concentrated in approximately six organizations, creating a new feudal structure where access to intelligence replaces access to land as the basis for extraction. The only path away from this new feudalism runs through genuinely open source AI, but the historical pattern suggests that technological revolutions deliver new forms of control rather than liberation. The outcome depends on whether distributed alternatives can be built before the window of opportunity closes.

Dawkins and Dennett’s evolutionary lens reveals the true danger of AI: not the machines, but human nature itself. With six corporations steering the human-AI merge, our ancient drives toward greed and tribalism make feudal capture nearly inevitable unless we build decentralized alternatives before the window closes.