A nice blog entry has been posted over on Deltoid, of ScienceBlogs [LINK]. It shows Hansen's 1988 climate model predictions of global warming along with observed global temperature. Despite how crude climate models were in 1988, Hansen's predictions are pretty much spot on. It is especially interesting to look at 1993, where the observations take a nosedive because of Mt. Pinatubo. They "recover" in about 2 years. Note that the credit on the figure is to the GISS page [LINK], but neither the blue line nor the extension of the red line (both observations) from 1998 to 2005 is on that page, and I don't know where that data come from. I tend to believe it though. If I find a better reference, I'll post it.
UPDATE: The red line (observations) actually isn't extended. Instead another dataset (blue) is just overlaid.