This position winds me up:
“I don't mean that it's important to consider the possibility that 100% of good contributions are LLM-assissted. I'm saying that we don't know the actual number and the blog post doesn't address it at all.”

This is a bullshit argument. When someone says the numbers aren’t good enough for them, can they articulate what numbers WOULD be good enough? What numbers would the project need to post in order to have their argument pass this person’s arbitrary muster? We can simulate numbers. We can imagine X contributions, Y% LLM, Z% good, etc. But this person wants the LLM-banning project to do all that work, publish the numbers to this person’s satisfaction, and take feedback.

I guess I’m saying if “the actual number” really makes the difference, the person can make a synthetic argument using dummy numbers to make the point. They can establish a methodology where all you have to do is put in the real numbers and a or pops out. They don’t. They just whine that the author didn’t do what they think they want.

Author: “our experience with triaging PRs pre-LLM ban does not match your assumption at all.”

Enthusiast: “I don't understand how you'd be able to tell.”

Never let someone else’s competence get in the way of your confidence.

But 90% of the top performers* are LLM users!

* Measures by LoC changes, I guess?