I think you would need a complicated set of metrics to claim something like "improved" that wasn't caveated to death. An immediate conflict being total number of articles vs impressions of articles labeled with POV biases. If both go up has the site improved?

I find I trust Wikipedia less these days, though still more than LLM output.

Care to provide any counter-examples? Please make it know if you end up using Wikipedia for your source of if Wikipedia's quality has changed

How in the world would you supply a counter-example for "the quality of Wikipedia has/hasn't improved substantially in the last decade"?

I also can't even read the second sentence. I think there are typos there, but there's no mental correction I can do to make it coherent for me.

I can't think of a better accidental metric than that!

I'll go ahead and speculate that the number of incoherent sentences per article has gone down substantially over the last decade, probably due to the relevant tooling getting better over the same period.

Know should be known