OpenAI and Google DeepMind demonstrated that their foundation models could outperform human coders — and win — showing that large language models (LLMs) can solve complex, previously unsolved ...