How did you measure it? Did you measure the efficiency that Intellisense, codegen, tools, etc.. provided to you before ChatGPT/copilot and what would be the delta?