OpenAI’s claimed Navier-Stokes solution faces questions about its physical realism

OpenAI said in September that agents had produced a solution to the Navier-Stokes Millennium Prize Problem, which concerns whether smooth three-dimensional fluid motion can break down. In an October reply, a critic argued that the solution meets the prize conditions through a forcing function that could never occur naturally, revealing nothing about how the equations behave under real-world physics. Another post asks whether that amounts to specification gaming.