If you suffer from a heart condition, you may want to take a dose of beta blockers before you watch this video of Paul Valdes, of the University of Bristol (who is, incidentally, Tamsin's boss). Or perhaps avoid it completely.
The bit on model assessment and the unknown unknowns is particularly amazing. One can't help feeling that someone should themselves get down to Bristol to explain the scientific method to the scientists down there. (One should note in Prof Valdes' defence his welcome admission that the models are tuned to achieve hindcast accuracy, although one can be equally appalled that he presents no other evidence to support the reliability of GCMs.)
The economics presented are rather extraordinary too.
Overall, I'm astonished that students would actually pay fees to listen to this stuff.