It's at least worth keeping in mind that this discussion thread is happening on a link to an article describing an occurrence -- the Laptev Sea not freezing by late October -- that aggregate data up through a couple of decades ago would not have predicted. Global warming itself is arguably a radical break from what aggregate data tells us about climate trends, and the (originally good faith) argument against it for decades was, essentially, that the "alarmists" were mistaking statistically insignificant temperature drifts for new statistically significant trends.

The theory that global warming is going to lead to food shortages in the relatively near future is not an out-of-the-blue flight of fancy on the part of random HN commenters; a UN panel in 2019 warned of this possibility, as did a 2020 IPCC report (pre-pandemic, no less), as did a 2016 study in the Lancet; this is just from a cursory examination of the Google results for "food shortage predictions global warming".

> Aggregate data is much more relevant than anecdotes.

Extrapolation of historic data trends isn’t always a good way to predict the future either.

Dinosaurs reigned for over 100 million years. Until they suddenly didn't.