> % of AI suggestions accepted vs edited this has to be the worst metric. anytime the llm wants me to read a diff of one file, im just gonna send it forward s

> % of AI suggestions accepted vs edited this has to be the worst metric. anytime the llm wants me to read a diff of one file, im just gonna send it forward so i can read the whole diff