Climate models are showing that a parameter (ECS aka “climate sensitivity”) has been increasing since 2019. This parameter determines how much temperature increase can be expected if atmospheric CO2 doubles.
and it turns out that by building in a strong preference towards the null hypothesis, science has tended to dramatically understate both the amount of emission, the amount they matter, and the feedback loops that can amplify this further.
like yea it has been simulated to death but everyone wanted to whistle past the graveyard and so those simulations had drastically optimistic assumptions built into them such that we are likely to crash through what was only a decade or two ago considered the “worst case scenario” by a decent sized multiple.
And frankly the root cause is really the people who go “I find this very confusing, don’t you have this all figured out by now? wasn’t it supposed to be global cooling in the 70s???”, the process was built around appeasing you so you didn’t dismiss the whole thing as alarmist and it turns out the process was built in a way that produced (unsurprisingly) over-optimistic results. The models missed low largely because royal-you wanted to play skeptic 30 years ago.
(And no, “global cooling” was never a thing and the idea that it was is more fossil-fuel propaganda and marketing. Even noaa is uncharacteristically blunt about this.) https://ams.confex.com/ams/pdfpapers/131047.pdf