The most advanced AI reasoning models collapse catastrophically at complexity thresholds — without knowing they have collapsed. They continue, confident and fluent, while the substance is gone. This is not a technical problem. It is a philosophical one. And Socrates saw the decisive limit 2,500 years ago: reasoning without the capacity to recognize its own failure is not reasoning. It is simulation. The mirror does not know that it is reflecting.

This image mattered to CP. It shows Federico da Montefeltro with his son. A renowned military leader during Renaissance. Ruler of a small and exposed principality. Surrounded by stronger powers. Trained for war. And yet the image does not show him fighting. It shows him reading. Investing in learning, in culture, in the life of the mind. For CP, the message was simple and uncompromising: Always be ready to fight. But never stop being open and curious. I worked with Claus-Peter Zeitinger for...

Some AI systems govern. Others reason. Behind this difference lie two competing ethical traditions — and two radically different assumptions about human judgment, power, and trust.