Until every last one of these people gets together and puts all their data into the world's biggest computer I'd take anything that is said with a pinch of salt. Or possibly a shovel full.
Actually, I'll take that a step further -- we need to be able to reliably predict global temperature over a time period, the more lengthy the better.

I had a 14 year career building computer models, and in the industry I worked in our models were REALLY, REALLY good at reproducing past events but were REALLY, REALLY BAD at predicting what would happen in the future.

A wise man once said "All models are wrong, but some models are useful". So the world's biggest computer and all the data isn't very useful if we don't know what we're doing. If we DO know what we're doing, we may only need a fraction of that data.

In any case, the only way to validate a computer model is by experimental validation -- use it to make predictions, than see those predictions come true. Which is notoriously tough in the case of climate because there are so many freaky variables. :)

Respectfully,

Brian P.