Comment by dag100

I think there was a major jump in AI capabilities from Anthropic and OpenAI between the end of 2025 and the start of 2026 that made them far more reliable at programming correctly. I wonder what changed in the secret sauce.