Skimming through, some of the ideas are quite scary. They'd make good starting points for "research gone wrong" scenarios encountered by Firewall operatives. As an example, the [url=http://www.darpa.mil/dso/thrusts/bio/biologically/neovision/index.htm]Ne... project[/url]. Imagine a research team makes breakthrough in imitating or even transplanting biological visual recognition ability, using an AGI as a test subject. Unfortunately, this subjects the AGI to optical illusions and subsequent hallucinations, and it is not used to or prepared for observing things that do not actually exist and is driven psychotic.