I think part of the problem is treating models as evidence rather than as hypotheses. Increasingly we see articles (and this is across a lot of disciplines, not

I think part of the problem is treating models as evidence rather than as hypotheses. Increasingly we see articles (and this is across a lot of disciplines, not just climate science) that are basically "we hypothesised X, and then we created a model that confirmed X". That's not how the Scientific Method works, folks. The models need to be built on a hypothesis, and then they need to make predictions, and then those predictions need to be proven by measurement of the actual phenomenon. That then gives some evidence (it's not complete proof) that the hypothesis isn't false. Instead, the current thinking seems to be that the models reflect reality and can be used to prove the hypothesis. The problem with this, of course, is that a model can be made to prove any hypothesis.