Due to the widespread idea that climate prediction has been predicting doom and gloom all the time, and has always been wrong about it, I have been entertaining myself by revisiting old climate predictions from 20 years ago. I can highly recommend it.

I mean, predicting the climate must be one of the hardest numerical simulations you can run. Climate models are famously incomplete. But the results turned out to be way more accurate than I would have thought. That's some impressive modeling, and I would assume that models have only improved since then.

The most chilling thing (and not in a good way) about it is that popular science type repotrs for the general public usually indicate an interval between a good case and a bad case, with a nice line in between. In retrospect we have been pretty much perfectly tracking the bad case line ever since.