Dyson is a physicist and mathematician, so his opinion on this matters exactly the same as yours - not a jot
So will you take my word as a computational physicist that climate models--which are nothing but computational physics done by climate scientists rather than computational physicists--are far too uncertain to be robust guides to public policy?
Because that is my professional opinion, and it happens that my profession is the one that matters when judging computational physics, whether it's done by climate scientists, Freeman Dyson, or anyone else.
I've read climate modelling papers. I've looked at climate modelling codes and there documentation (mostly AR4, which is somewhat out of date now.) I was appalled by what I saw: it's all a good attempt to work things out, there's nothing wrong with it as science at all, but I'd rather use Wall Street financial models to guide public spending policy than climate models to guide climate policy. They have a much greater chance of being accurate.
This is not to say that climate models aren't useful inputs to the policy debate, but their accuracy if fantastically over-estimated by policymakers. GCMs have gotten Arctic warming badly wrong (the Arctic has warmed much faster than anyone anticipated) and missed the current--likely temporary--flattening of "global average temperature" increase. This is no surprise you a) look at the models and b) have the professional competency of a computational physicist to judge them. They just don't do the things that accurate models integrated over long timescales have to do, like conserve mass and energy natively.
Models before around 2005 were especially bad with energy conservation, fixing it up by redistributing energy across cells after each time step. Climate scientists were apparently OK with that, because they didn't know enough computational physics. Anyone who has spent a career building models that eventually get checked against reality knows that that is a virtual guarantee that the result will be unphysical nonsense. This is not a political statement: it is simply a fact.
So by all means dis Dyson for not being a climate scientist. But since GCMs are computational physics, you must take my word as a computational physicist over climate scientists, or admit you really don't care who is saying what so long as they say what you agree with.