I saw that tweet as well and was thoroughly confused that it was being treated as “definite proof” - why does he think he can draw any conclusions about long-term trends by comparing 2 datasets that are only 2 years apart?

Based on that logic I could easily prove that house prices are going down (“this particular house was worth $X in 2017 but less than that in 2019”), global warming doesn’t exist (“in one part of the world 1st June 2017 was hotter than 1st June 2019”) or pretty much any ridiculous thing I wanted. Are people really this ignorant about how statistics work?