OpenAI mistranslated mathematics into code for its Navier-Stokes proof
OpenAI’s recent attempt to encode a Navier‑Stokes proof into software encountered a critical error when the system incorrectly translated mathematical expressions into code. The mismatch caused the generated program to diverge from the intended proof steps, highlighting challenges in automating advanced mathematics. Researchers note the incident underscores limits of current AI code‑generation for complex theoretical work.