Cardano founder Charles Hoskinson said artificial intelligence has advanced in advanced mathematics far beyond his expectations, particularly in generating and formally verifying complex proofs. In a recent YouTube broadcast, he discussed claims that OpenAI had produced an AI-generated approach to the Navier-Stokes existence and smoothness problem, one of the Clay Mathematics Institute’s Millennium Prize Problems. The problem asks whether sufficiently well-behaved solutions always exist for equations describing three-dimensional fluid motion, with implications for mathematics, physics, fluid dynamics, aerospace engineering and mechanical engineering. Hoskinson, who previously founded a center for formal mathematics at Carnegie Mellon University, said he had expected formal systems mainly to improve collaboration among mathematicians rather than write proofs themselves. He also questioned the provenance of the work and warned that ideas shared with frontier AI models in the cloud may no longer remain private. He used the episode to argue for private AI environments that protect research logs and intellectual property, while describing the system’s ability to improve, iterate and formalize existing work as strikingly humanlike.