I wasn’t trying to build Artificial General Intelligence. I was trying to fix a small problem in how computers think — the randomness, the drift, the fact that identical code could behave differently on two machines. What started as a search for precision turned into a framework for verifiable thought. By making computation deterministic and reproducible, I realized I’d built the foundation for intelligence itself — not artificial, but synthetic, predictable, and provable.