For example, if a new rootkit starts spreading and reaches a malware honeypot, any skilled security specialist could get a memory dump of the system, find the malicious code and reverse engineer it, right? So what prevents known malware from being readily neutralized once they are launched?
I think the easiest way to answer your question is that most malware is coded using Polymorphic Code and the majority of Anti-Virus still uses signature based detection.
The reality is that a good chunk of the population does not keep their software up-to-date which means the original attack vector could potentially be re-exploited with a different payload. A lot of people put a lot of thought and effort into prevention, however not many put it into detection so once a box has been owned there's usually an avenue to re-exploit.