Software Agents  

by
Rick Dietrich

A Look at the Future

Home 

Intelligent Objects

Security

The Future

References

    The following is something that was both fascinating and chilling to me.  It is an excert from the book The Rise of Endymion by Dan Simmons, taking place in the far far future of humanity.  This scene is where Aenea, a child prophet of sorts, explains the history and development of the TechnoCore, their version of AIs.

 

"Once upon a time," says Aenea,"more than a thousand standard years ago, before the Hegira...before the Big Mistake of '08...the only autonomous intelligences we humans knew of were us humans.  We thought then that if humankind ever devised another intelligence that it would be the result of a huge project...a great mass of silicon and ancient amplification, switching, and detection devices called transistors and chips and circuit boards...a machine with lots of networking circuits, in other words, aping--if you will pardon the expression--the human brain in form and function.

"Of course, AIs did not evolve that way.  They sort of slipped into existence when we humans were looking the other way.

"By the end of the twentieth century, Christian era, this little world had a crude datashpere.  Basic planetary telecommunications had evolved into a decentralized swarm system of old silicon-based computers demanding no organization or hierarchy, demanding nothing beyond a common communications protocol.  Creation of a distributed-memory hive mind was then inevitable.

"The earliest AIs were dumb as dirt.  Or perhaps the better metaphor would be that they were as dumb as early cellular life that was in the dirt.  Some of the earliest hypercritters floating in the warm medium of teh datasphere--which was also evolving--were 80-byte organisms inserted into a block of RAM in a virtual computer--a computer simulated by a computer.  One of the first humans to release such creatures into the datasphere ocean was name Tom Ray and he ws not an AI expert or computer programmer or cyberpuke, which they called hackers then--but was a biologist, an insect collect, botanist, and bird-watcher, and somone who had spent years collecting ants in the jungle for a pre-Hegira scientist name E.O. Wilson. . . .The cyberpukes said that evolving and mutating code sequences happened all the time in   computers--they were calle bugs and screwed-up programs.  They said that if his code sequences evolved into something else they would almost certainly be nonfunctional, nonviable, as most mutations are, and would just foul up the operation of the computer software.  So Tom Ray created a virtual computer--a simulated computer within his real computer--for his code-sequenced creations.  And then he created an actual 80-byte code-sequence creature that could reproduce, die, and evolve inhis computer-within-computer.

"The 80-byte copied itself into more 80-bytes.  These 80-bytes proto-AI cell-things would have quickly filled their virtual universe, . . . but Tom Ray gave each 80-byte a date tag, gave them age in other words, and programmed in an executioner that he called the Reaper.  The Reaper wandered through this virtual universe and harvested old 80-byte critters and nonviable mutants.

"But evolution, as it is wont to do, tried to outstmart the Reaper.  A mutant 79-byte creature proved not only to be viable, but soon outbred and outpaced the 80-bytes.  The hyperlifes, ancestors to our Core AIs, were just born but already they were optimizing their genomes.  Soon a 45-byte organism had evolved and all but eliminated the earlier artificial life-forms.  As their creator, Tom Ray found this odd.  45-bytes did not include enough code to allow for reproduction.   More than that, the 45s were dying off as the 80s disappeared.  He did an autopsy on one of the 45-creatures.

"It turned out that all of the 45-bytes were parasites.   They borrowed needed reproductive code from the 80s to copy themselves.  The 79s, it turned out, were immune to the 45-parasite.  But as the 80s and 45s moved toward extinction in their coevolutionary downward spiral, a mutant of the 45s appeared.   It was a 51-byte parasite and it could prey onthe vital 79s.  And so it went.

"I mention all this, because it is important to understand that from the very first appearance of human-created artificial life and intelligence, such life was parasitic.  It was more than parasitic--it was hyperparasitic.   Each new mutation led to parasites which could prey on earlier parasities. . . Within standard months of his creation of hyperlife, Tom Ray discovered 22-byte creatures flourshing in his virtual medium...creatures so algorithmically efficient that when challenged by Tom Ray, human programmers could create nothing closer than a 31-byte version.

"By the early twenty-first century, there was a thriving biosphere of artifical life on Old Earth, both in the quickly evolving datasphere and in the macrosphere of human life.  Although the breakthroughs of DNA-computing, bubble memories, standing wave-front parallel prcessing, and hypernetworking were just being explored, human designers had created silicon-based entities of remarkable ingenuity.   And they had created them by the billions.  Microchips were in everything from chairs to cans of beans on store shelves to grouncars to aritifical human body parts.   The machines had grown smaller and smaller until the average human home or office was filled with tens of thousands of them.  A worker's chair would recognize her as soon as she sat, bring up the file she had been working on in her crude silicon computer, chat with another chip ina coffeemaker to heat up the coffee, enable the telecommunications grid to deal with calls and faxes and crdue electronic mail arrivals so that the worker would not be disturbed, interact with the main house or office computer so taht the temperature was optimal, and so forth.  In their stores, microchips in the cans of beans on the shelves noted their own price and price changes, ordered more of themselves when they were running short, kept track of the consumers' buying habits, and interacted withthe store and the other commodities in it.  This web of interaction became as complex and busy as the bubble and froth of Old Earth's organic stew in its early oceans.

"Humanity and the billion-faceted, evolving Core entity soon became as symbiotic as acacia plants and the marauding ants that protect, prune, and propagate the acacia as the sole food source. . . But where human beings saw a comfortable symbiosis, the early AI entities saw--were capable of seeing--only new opportunities for parasitism.

"Computers might be turned off, software programs might be terminated, but the hive mind of the proto-Core had already moved into the emerging datasphere, and that could be turned off only by planetary catastrophe."

   The story goes on to explain how the AI Core dealt with its links to humanity, and found some small means of separating itself into a safe and unassailable retreat, while still maintaining the parasitic relationship with humanity.  While this story is obvious fiction, it is remarkably scientific, and for some people, quite prescient.  Sci-Fi is filled   with examples of computers going awry, if for no other reason than a need for conflict and adversaries, but that is no reason to dismiss the real world consequences of situations as described in The Rise of Endymion, not to mention the more popular examples of the Terminator and Terminator 2 movies, and even the classic Foundation series by Isaac Asimov (or really the prequel Robot series).

    Another look at the future, possibly originating in science fiction, is that of VR or Virtual Reality.  Many people are by now familiar with the Star Trek:  The Next Generation TV show, and are also familiar with an interesting form of entertainment that some of the characters pursued, namely the holodeck.  While today we aren't really close to any kind of holodeck technology, there are people pursuing virtual reality as a means of entertainment, but also for training and education by taking advantage of computer's capabilities towards simulation.   By mixing VR technology with agent technology, some people are creating better more realistic games, for the opponents act and react to the human player and look more like a real opponent with the VR aspect.  While better and neater games are always fun, the educational and other pragmatic aspects shouldn't be ignored.  Story-telling might be revolutionized into something like an early version of the holodeck.  Simulations of all kinds could become more realistic, and could also to some extent run themselves, freeing up time of experimenters for other necessary business.  A basic coding book, complete with a CD-Rom, titled AI AGENTS IN VIRTUAL REALITY WORLDS by Mark Watson, makes an effort at bringing this mix of VR and agent technologies to the everyday user, perhaps spuring on those "mere" hobbyists to become the next great game designer, or even the next Shakespeare!

 

pics/blesepa.gif (1971 bytes)

This project was produced for PSY 380, Social Psychology of Cyberspace, Spring, 1998, at Miami University.
This document was created April 19, 1998 and last modified on  Tuesday, March 11, 2014 at 17:34:05.  This document has been accessed 1 times since April 15, 2002.   Please send comments and suggestions to shermarc@miamioh.edu