Blogging the Noosphere

A few people have been curious about the meaning of the word metacortex after seeing it here on the blog. I saw a column today by David Brooks on the New York Times that probably does as good a job as anything in giving an intuitive feel of the evolving concept of an exocortex.

Vinge, the originator of at least the term Singularity, if not the concept itself, has proposed 3 main routes to a technological singularity. One route is by the development of a smarter-than-human artificial intelligence. A second route is by the genetic enhancement of human intelligence to above current human maximum. But a third route Vinge proposes is through human augmentation, a man-machine cybernetic merger. In the NYT article I referred to earlier, you can get a feel for what he is talking about, of how technologies such as GPS auto navigation systems along with calculators, iPods, BlackBerries and other devices provide many of the functions of a cortical memory in an externally stored and accessed form. Now I don’t know if a technological singularity is really in our future or not, but one must admit that people do have a point with these observations of the trends we are seeing today. Our information technology devices and systems are greatly enhancing our memory capacities and practical functionality in many interesting and unforeseen ways. And this is happening today, right now, and it is not our children but we who are participating in this transformation.

Metacortices and Exocortices (from Wikipedia ): Though etymologically similar, the two are quite different — the former, in effect, being composed in various ways of the latter. An EXOcortex can best be described as the portion of a trans- or posthuman entity’s brain (or cortex) which exists outside of that entity’s primary computing structure, usually the brain inhabiting a person’s “meatbody.” For example, a person’s exocortex could very well be composed of all the external memory modules, processor, and devices that the person’s biological brain interacts with on a realtime basis, thereby in effect making those external devices a functional part of the individual’s “mind.” A METAcortex, on the other hand, is a processing construction built entirely out of the connections between other processing constructions — a sort of “higher brain” composed of lesser brains, all of which contribute to its functionality.

The 2007 Hugo Award Winners were announced awhile ago and I saw that Vinge had won another Hugo for his excellent novel Rainbows End. I love Vinge and read everything he puts out, as soon as it comes out. He is one of the few I will pick up in hardback, which I did with Rainbows End.

One thought after rereading. We are currently in the middle of our Fall 2007 AI Course. Vinge is a Computer Scientist by training, and it shows in his work. The ideas just leap off of every page of this novel. I especially love his concepts of Analyst Pools. These are really a bit of a fleshed out version of a borganism (a term invented by Stross, I believe). I can almost see how a system like this might be created in today’s world of Social Networking Systems and Wisdom of Crowds organization. To bring it back to our AI class, imagine a big A* or min-max search specification being conducted in real-time. But instead of machine generated and evaluated heuristic functions, imagine analysts instead providing human-level-intelligent type heuristic estimation of states, possibly mitigated by a Prediction Market to allow many competing heuristic estimates of states to be combined. I could imagine, if you could build the software support system to manage 10s or 100s of thousands of analysts all working collectively on such a machine/human guided search, you would have something very like what Vinge envisions in Rainbows End. I’ll just say that I know of research initiatives at NSF and elsewhere that are envisioning and working towards the possibilities of such real-time social borganism as this.

I’ve been doing a bit of planning for my Introduction to Artificial Intelligence class, to be taught this fall. For those who haven’t yet discovered, MIT offers a wonderful resource in their MIT OpenCourseWare. Basically they have opened up many of their courses materials for open and free use. A great resource for educators and self-learners alike, as they state on the site.

Anyway, I think I will be using one of my favorite texts, and the one I used as a student for my AI course (though it has been updated in a second edition), Russel and Norvig’s Artificial Intelligence: A Modern Approach. In my opinion, a very good introduction to the basics, with as the title suggests a modern viewpoint cognizant of the successes, failures, strengths and weaknesses of the standard techniques and algorithms of AI.