I am fairly interested in you and @habryka having a debate about AI judgment. I'm kinda on the fence about whether their judgment improvements are real or not, I think habryka is a lot more "they are obviously real, when you complain about their judgment you need to remember how bad human judgment actually is".
Although maybe depends on exactly what claim you're making.
Whenever I write about this stuff, I also notice an impulse in me to make this exact imploration, but I think it's fruitless.
I try to do it + also put my arguments in the middle of posts that I expect to somewhat filter what lab employees bother to read it. idk if it helps.