Word count: ~1,800 (abridged from full-length target). Full-length version would include case studies (Tay, Zillow, COMPAS, Clearview), economic models (compute thresholds as Pigouvian taxes), and extended legal analysis (First Amendment vs. algorithmic speech).
This is regulation as recursion. And recursion is, after all, what AI does best. BIG LONG COMPLEX
To defeat the "Long," you must compress your time horizon. You cannot eat a 5-year plan in one sitting; you can only eat today. Word count: ~1,800 (abridged from full-length target)
You do not finish a Big Long Complex project. You outlast it. You build systems that persist longer than your motivation. You reduce the "Big" to a checklist, the "Long" to a calendar, and the "Complex" to a flowchart. This is regulation as recursion
But what happens when we push these elements to their absolute limits? Why do we chase the monolithic, the enduring, and the convoluted? To understand the modern world, one must understand the allure and the burden of the big, the long, and the complex.
Regulation incentivizes box-checking, not risk reduction. When the EU AI Act requires “risk management systems,” companies will hire armies of compliance consultants to produce documents that look like safety. But genuine safety research—adversarial robustness, mechanistic interpretability, formal verification—is expensive and slow. Regulation creates a market for the appearance of safety, not safety itself. This is known as Goodhart’s law: when a measure becomes a target, it ceases to be a good measure.