The arguments I always heard back in the 90's were how we didn't know for sure, but we had to act now and couldn't take the risk of being wrong. That morphed into "the science is settled" in the 2000's. It has taken time for the real science to set the record straight.
That was certainly one of the big problems when the new models were introduced. You had people like Michael Mann (climate scientist wannabe) cherry pick his data to prove his hypothesis (versus disprove the null), and people like James Hanson, who was at one time a reputable climate scientist, go off the deep end bringing politics into science.
Those are a couple of overt examples. Taking the model Mann used though, someone else could have cherry picked data to prove the opposite. What I think Dyson was partially referring to though, was that if you took that same model, and input all available historical climate data in an unbiased way with adequate QA/QC, that was where you had to be open-minded on what the model may or may not indicate.
Ten years later we're getting some good data on what they're not indicating, but we now have so-called "climate scientists" that aren't accepting the failure of their hypothesis for the models. This is the problem with the AGW believers - they have to be "right" when that's not what science is about. Failure of a model is just as important a discovery as it would be if the models were accurate. Failure in science is not really failure - it's success at narrowing possibilities. The AGW people won't accept that - they have to "win".
"Consensus science" is not science - it's politics.