Comment by gojomo

Thanks, that's great!

But: if all developers did 136 AI-assisted issues, why only analyze excluding the 1st 8, rather than, say, the first 68 (half)?