Chain-of-thought prompting gives superior responses in LLMs. OpenAI should have implemented that a while ago. That's not "cheating" – that's programming.

→ View original post on X — @aibreakfast