[holtkam2]

I’m convinced this is the root cause of the strange phenomenon I call “LLM coding assistants don’t increase team velocity in the long run in real world settings”