To me, being able to intelligently anticipate and answer questions about "the next version of the primary tool of my trade" is part of what I'm being paid for. If my CIO ever asks, "Would there be security improvements if we upgrade to the newest version?", I want to be able to answer that, and not have to say, "Well, I haven't bothered to look into that because our company is too cheap to ever upgrade".
Valid point if security was the primary concern, but it's not the only reason to upgrade. I keep an eye on the latest released and real versions for features of value and merit. Eventually I will either support them or be programming for them. The industry I work in is very conservative and concerned with ROI, so filling Redmonds coffers every two years by upgrading to every single release isn't a priority. This has paid off by skipping duds such as Windows Vista, waiting for various SP1 and SP2, dodging late bugs, etc... Having been in this business since the DOS days, it pays to let the others take the arrows and glean information from their experiences when the overall cost for licenses/training is high. Not even the developer copies in MSDN tempt me much anymore. Also in a smaller shop/company, the cost of failure is higher and time is dearer.