I would like to add one more entry to the long list of biologically or reality inspired, ALife-like architectures (or at least architectural directions and decisions): a cyberagent that primarily and constantly risk-manages its own existence and tries to survive. I will remind you that, according to contemporary biology, most (if not all) of the diversity and adaptations among species, including cognitive adaptations, are a result of reproductive competition - if the brighter gene/specimen has even a minuscule reproductive advantage over its peers, in the long run it becomes dominant with mathematical certainty. It is hard to argue with this reasoning, but it is good to point out that survival still outweighs reproduction as an imperative, if you cannot make it through infancy and childhood and if you die to soon as a "parent" then your reproductive advantages are as good as nada.
Psychosomatic considerations aside, we could divide an organism's risk management as conscious and unconscious, the latter corresponding perhaps to the immune system and all kinds of resilience built into plants and animals so that they do not die at the first hurdle. Still, consciousness and culture are ever expanding into the unconscious with a variety of feedback loops, we try to boost our immune system, move to safe neighborhoods, filter our water etc. In the same fashion, it would be desirable for a cyberagent not model risks at its "own" level of abstraction, for example by measuring the performance of its submodules under certain circumstances, but to incorporate risks such as power being unplugged, virus infection, thermonuclear war etc. I would not like to assert that human cognition developed to model increasingly accurately a rather convoluted "attack surface" of the human organism. But I find it infinitely more likely that the questioning and myth making human mind initially and for the longest time managed risks in its environment, and lately and shortly tackled gravity and gravitons. Obviously a very active imagination will be key to risk management, and as an added bonus we may get a cyberagent that will not "die" when we format the hard disk and unplug the computer - uh oh! AT ------------------------------------------- AGI Archives: https://www.listbox.com/member/archive/303/=now RSS Feed: https://www.listbox.com/member/archive/rss/303/21088071-c97d2393 Modify Your Subscription: https://www.listbox.com/member/?member_id=21088071&id_secret=21088071-2484a968 Powered by Listbox: http://www.listbox.com
