Ladybird’s blog post is arguing that it used to be tolerable to spend the amount of time evaluating this for every PR, now it’s just unmaintainable for their effort-time.
How do you do that efficiently so it can scale?
