I was really quite surprised that they'd parade around 722 manuscripts with only less than half (or really, close to only a third) of it being machine-checked. It's genuinely equivalent to the overconfident and blind drive-by PRs in software that people here should be well familiar with, and those are not exactly the hallmark of quality usually. Quite the curiously dangerous game to play when you're trying to scoop an entire field of study on an industrial scale as powerplay.

Or it's a touching show of openness. They're not in this for the ego-boost, they're sharing data on how things are working. Naive of them to expect gratitude, I guess.