I think one of many issues with the car windshield observation is we don't have a good baseline. So we notice that in 1990 there were more bugs on our windshields compared to 2020 after a drive through Kansas. But maybe 2020's insect levels are actually a return to normal, preindustrial levels after an anomalous century of high bug populations? We don't have a bug windshield strike stats from 1890.

Honestly it's such a dirty metric with so many compounding variables I'm shocked that anybody is seriously considering it as a meaningful data point beyond the most basic, surface-level understanding of the issue. To be honest the whole thing reeks of the usual hacker-news crowd issue where people want to pretend they know about a topic (e.g. systemic issues in pollinator/agricultural pest population management in high-intensity modern industrial agriculture) but are underestimating the complexity of the issues by a huge magnitude.