Check out Zenpundit - Super-Empowered Individuals
"This post is one of the natural extensions of what we have been discussing. I don't think there is any doubt that it is inevitable. I suppose the 'when' depends on what system is perturbed/attacked. It will be done as our understanding of network theory and complexity advance; to have, say, an individual do tremendous damage, that person will need the means of mapping out and understanding the levels of connectivity inherent to the system, whether that system is social, electronic, environmental, industrial, etc. Even with a lack of understanding of the system's multi-dimensional topology in whatever relevant phase space, I can imagine someone developing and using one of these newer adaptive genetic computer algorithms...this type of program can 'learn' as it crunches data, and can adapt itself to the system. It is along the lines of the programming being tried for intelligent robots, etc. That is probably the scariest scenario to me."
My thinking is that at some point, as these types of algorithms and technology further develop and become more widespread, cheaper, and user-friendly, it will no longer take an expert in the relevant fields to do damage to different systems of concern...some amateur hacker type can just unleash a virus built around such software and the software will be 'intelligent' enough to do damage on its own. As all aspects of life become computerized at some level, this form of super-empowerment is, in my mind at least, the single greatest technology security issue that faces us in the future (and is on a level near that or arguably equivalent to nuclear and biological terrorism...while it may not cause immediate death and physical destruction, the potential to adversely affect countless millions of people is there). Resilience in all computer systems and networks is absolutely essential.
Zen, good job as always, my friend.