From the more-bad-news-for-those-who-want-to-live-forever department: http://pipeline.corante.com/archives/2010/10/08/does_resveratrol_really_work_and_does_srt1720.php

The current study dials that back to levels that could be reached in human dosing. What they saw was no effect on lifespan at 0.5 micromolar, which would be a realistic blood level for humans. When they turned up the concentration to 5 micromolar, there was a slight but apparently real effect of just under 4%. Now, 5 micromolar is a pretty heroic level of resveratrol - I think you could hit that as a peak concentration, but surely not hold it.

Nor is the bad news just for resveratrol:

Oh, and there's another interesting part to this paper. The authors also looked at SRT1720, the resveratrol follow-up from Sirtris that has been the subject of all kinds of arguing in the recent literature. This compound is supposed to be several hundred times more potent than resveratrol itself at SIRT1, although if you've been following the story, you'll know that those numbers are widely believed to be artifacts of the assay conditions. And sure enough, the authors saw no effect on C. elegans lifespan when dosing with physiological concentrations of SRT1720.