On Wed, Feb 18, 2015 at 08:21:28AM +0200, Nanograte Knowledge Technologies via AGI wrote: > @ Logan > > Your point on diversity is well taken. > > We are debating using words, so semantics should be incorporated accordingly. > I think any debate about intelligence, on an AGI forum, is quite relevant sure >. However, your post seemingly raised the question whether mutation should be >regarded the same as replication. I think they are vastly different in >complexity, perhaps even in as far as replication (in the sense of procedural >copying) being on a complexity scale of 1 vs mutation being a 10. > well there is at least one species of lizards which clone themselves. they still go through a mating process to stimulate ovulation, but it is amongst females.
the mutation is certainly more difficult, at it requires searching for and copulating with a comptible mate, usually through some ritual. Clones replicate much faster, such as a notorious cloning crayfish. Though as you know are more suceptible to die-off, due to lack of diversity. In terms of robots, I imagine it as a female robot queen, much like an ant, say on Mercury, or in Antarctica. She would have a mine and sister workers that maintain the energy harvesters, smelters, and manufacturing. The males would venture and explore the world, getting new ideas and insights for how to improve processes back home. Similar to how it works with many animal groups, like sperm whales. The fit males which survive the journey, and find a willing mate can reseed. Perhaps merging their knowledge with the queens. > Perhaps, a lizard's tail assuming its previous form would be considered > replication, would you not say so? In other words, does a lizard, some > variants of which can reflexively drop tails, think about regrowing the tail, > or is there a chemical it is quite unaware of, which autonomously triggers > regrowth? animals/phones are second-density organisms in the "Law of One" view. Thus they are desire based, where they fulfill various desires in different ways. For a lizard it may be to get and eat some food, for a smartphone it may be to retrieve and render a web page. in terms of a tail falling off, it's similar to an x-server crashing, and then being restarted by the desktop manger. >Is an instinctive (designed) process a consequence, and/or evidence of local >thought, or something else, and how is that different from behaviour, which >would seemingly always require some form of contextual awareness? smart phones feature a rich variety of behaviours, much like lizards. > > What is this chip you mentioned? It's actually a prospective fabless semi-conductor company. > What does it do and how would it be useful for AGI? It focuses on producing libre (open-source) chips. It would be useful as AGI's could help design them. In terms of a particular chip good for AGI, the best I know of right now is the Zynq-7000, CPU+FPGA it is found on the Parallela boards commercially. I'll get a Parallela board to start testing sometime in next few months hopefully. Though still have to get my Odroid OpenCL GPU working, seems like it needs one of their HDMI monitors to function. Can't seem to find headless settings which work. Unfortunately Zynq is proprietary, and needs proprietary toolchains to get from an HDL program to circuitry. A Libre FPGA would ideally allow an AGI to customize the FPGA while it "sleeps/dreams". During the day it could make note of places where improvements can be made, much as human brain does by setting calcium bonds. Then during deep sleep these new neurons are formed, and tested in dreams. The HDL testbench tests the design by flooding the circuits with possible data, and then it is ready for implementation on the FPGA, where it could be tested again before waking for new day. A program moved from a pure software implementation, into an FPGA, can experience over 4 orders of magnitude in speed-up and almost as much in energy savings. so that's ~10,000 times faster and more energy efficient. > Can it be printed by a 3D printer? Ideally, once RepRap or it's descendants have a printing resolution of a micron or measured in nanometers, then yes. > > Rob ------------------------------------------- AGI Archives: https://www.listbox.com/member/archive/303/=now RSS Feed: https://www.listbox.com/member/archive/rss/303/21088071-f452e424 Modify Your Subscription: https://www.listbox.com/member/?member_id=21088071&id_secret=21088071-58d57657 Powered by Listbox: http://www.listbox.com
