I wonder how did they do that without predicting anything about CO2 usage in those 4 decades? You need to realize there are two questions: 1. What will be the

I wonder how did they do that without predicting anything about CO2 usage in those 4 decades? You need to realize there are two questions: 1. What will be the climate decades into the future? T(CO2, t)=? 2. What is the effect of CO2 on all of this? T(CO2+ΔCO2, t)=? Getting zeroth order of T is one thing. They did it pretty badly anyway. Getting the first order of T with respect to CO2 is another thing entirely. Getting the answer to 2 correctly is much, much harder. You want to do the the derivative by CO2 ppm, and as you do that, you have 10^-5 factors on most things, because that's how much CO2 ppm there is. I don't buy the narrative that they did this difficult thing correctly. Supposedly the leading order is something like how much CO2 increases absorption. It's going to have this 10^-5 factor of CO2 ppm there, and it's going to be what's causing this whole thing to happen. But if there are greater than 10^-5 errors in this calculation, the whole thing is useless. And they don't actually do this analytically, they just run it many times in a simulation. I just don't buy the narrative that they manage to get meaningful signal out of that, not with how their code looks like, how chaotic the real system is, and their uncertainties about other parameters, there's just no way in hell they know this correctly. Their evidence is that they have some pretty rough estimates for the zeroth order. The evidence for their confidence of the first order is extremely weak. And the narrative that CO2 matters is only meaningful if they have a good estimate of the first order.