What the paper in question really says...
Having now skimmed through the original paper (thanks for the link, Simon!), below is my explanation of what's going on. It's not surprising that Simon and others here have misunderstood the paper, as the climate system is quite complex and our current knowledge of past climate more like peering through foggy muck rather than reading a book in a brightly lit room.
One of the pesky issues that bedevils those who study the history of our atmosphere over geologic time periods is that when global temperatures change significantly, as far as has been determined, this always seems to happen hundreds or thousands of years *before* rather than *after* CO2 levels change, making a simplistic CO2-caused-temp-change mechanism a hard sell, at least to the general public.
In this paper, they (a) use CO2 data from high-deposition core sites (thicker ice per year means smaller time error bars) and (b) infer Antarctic temperature data from a group of cores around Antarctica rather than a single core from just one region. They claim that this approach allows them (using fancy maths) to correlate the CO2 levels and Antarctic temperatures more accurately and with smaller predicted errors than has previously been done.
Their result: changes in CO2 came about 160 years *after* changes in Antarctic region temperatures, plus or minus about 220 years. This is shorter than previously thought. Since 160 years - 220 years is -60 years, their results suggest a small but significant chance that CO2 changes actually happened a few decades *before* the temperature changes.
What does this mean for the current controversy? Not much, directly. However, it places some constraints on global climate models that are different from those previously thought to apply, which in the best of all possible worlds would be quickly used to improve those models and better understand the details of how and why the last ice age ended.