The length of time between the development of security patches and the development of exploits targeting the security holes they address has been dropping for some time.
Hackers exploit this period of time - the so-called patch window - to launch attacks against unpatched machines. Typically, exploits are developed by skilled hackers versed in the arcane intricacies of reverse engineering.
However, hackers have now begun using off-the-shelf tools to at least partially automate this process, a development that might lead to exploits coming out hours instead of days after the publication of patches.
Security researchers at Berkeley, the University of Pittsburgh, and Carnegie Mellon have launched a research project investigating the approach (pdf.), which relies on comparing the configuration of patched and unpatched machines.
And, a very good response from Roger's blog: