HOME | POLITICS | SPORTS | LIFE | SCI/TECH | OPEDS | HELPFUL TIPS
|
Nov. 7, 2005 This is the first part of a two part column dealing with artificial life. First up, “machine life”, or, rather, non-biological life. I say this because technically, biological life are machines. Imagine a very large room with piles and piles of LEGOS. Now imagine that a LEGO robot was placed into this room. This robot was constructed using the LEGO Mindstorm kit. That kit includes blocks that are computers, blocks that are sensors and, when combined with LEGO motors and gears, can make some rather sophisticated robots. Perhaps this robot has a central, rectangular, body, with two big arms at either end. The arms are each bigger than the body, and the robot has no top or bottom – it can lay on any of the four sides of the central body. Now imagine such a machine moving itself around slowly using its two arms – not by walking, but by pushing or pulling or spiraling itself along the smooth floor. It's looking for LEGOS. Following its program, it finds the right pieces, manipulates them, and sticks them together. It's making copies of itself. Once a copy is made, it transfers its own program into its copy, which then pops to “life” and crawls along, slowly making copies of itself as well. After a while, the whole room would be full of these robots and depleted piles of LEGOS. A nice bit of fiction, but we don't have any such system yet. For starters, the Mindstorm system is limited in processing power – we'd need to communicate via radio with a PC for the brains required to examine and find the right blocks. Further, the sensors aren't sophisticated enough to detect the right blocks, and it would be rather tricky making a LEGO arm that could be both delicate and strong enough to assemble LEGOS. We run into a complex situation here – if the arms are too complex, then it requires a more complex machine to assemble them, but that which is assembling them is also made of LEGOS – you can see the loop here. It's a very difficult problem, but if solved would have vast rewards. It would be the first basic steps toward artificial life of a kind – self-replication is not only profound but useful. In the real world, there aren't piles of LEGOS around, but we do have rocks and gases and dirt and if we dig, useful minerals. NASA seriously considered such an idea for a self-assembling moon base. Rather than spend billions sending all manner of sophisticated probes to the moon, it planned to send units that would use the materials found on the moon to make copies of worker vehicles. Imagine the costs savings, if you could construct your own work force from materials laying about! The NASA plan envisioned solar powered furnaces melting and molding lunar regolith (dirt) into the parts required for these vehicles. The brains of these machines would be sent as “vitamins” from Earth – that is, this wouldn't be true self-replication – the machines would not be autonomous. They would be controlled from the Earth, and would not make their own memory chips and computers. The big cost associated with space exploration is pulling material out of this deep gravity well we find ourselves at the bottom of, and the less we need to send, the better. Even pseudo 'clanking replicators', or Von Neumann machines or 'universal constructors', as they are sometimes called, would be a monumental step forward. Let's imagine for a moment a true clanking replicator. Such a machine would be able to find energy sources and raw materials on its own and make copies of itself. Perhaps a probe could hit an icy moon and use the water to get hydrogen, and whatever other minerals found there, to make the copy and fuel both itself and its “offspring”. The computer code running the craft could be considered DNA. If enough of these things run around and made copies of themselves, we could imagine a sort of evolution taking place, as transcription errors take place. If we ever reach this stage, we will have truly created artificial life – all of the requirements are met: they consume, they excrete, they move around, they are complex, they reproduce, they evolve, and so on. Right now, we have very crude attempts at this, some involving LEGOS, but not nearly at the complexity envisioned above. We also have mostly autonomous factories that can even order required parts on their own. Humans deliver the parts, and the factories don't make other factories, but it's probably as close as we've yet come. Another similar idea is within reach. It goes back to the LEGO theme. It consists of modules, each with a specific function. The “machine” isn't assembled, but sent, lets say, to Mars, as a “box” of modules. There, depending on conditions, the pieces assemble or are assembled by a factory to whatever is required. In other words, imagine little cubes with computer chips and wireless network communications ability. Some have wheels, some can extend a leg (like a TV antenna), some can twist and rotate, some are sensors. If each module can also handle basic movement up and around other modules, then we can have a dynamic system that changes to the environment. In other words, if there is a flat area, the wheeled modules can climb to the bottom and reattach so the whole can move faster, and if we need to descend into a crater, the modules can assemble into a crawling or snake like form. Some modules would collect either solar energy or act as batteries and as the modules connect, they can exchange power. Such a system could either consist of very many specific, one function modules or fewer, more general purpose, modules. Because each module would ideally have some processing power, the unit could also break up into smaller units and corporate at certain tasks or go off in entirely different directions doing separate tasks. Interesting work is already going on in this field, both with generic modules as well as machine cooperation – I've seen some rather impressive examples of the latter: robots cooperating to move a couch around a living room, and a bunch of small simple robots cooperating to play a simplified version of soccer. Both ideas, replicators or modular “build on demand” vehicles, have redundancy as a main benefit. Gone could be the days of the single, ridiculously expensive, probe, that, if even slightly damaged, seriously affects or cancels the mission. All kinds of hybrid system can be imagined, where the replication system is able to find and use raw materials to replicate just some of the simpler modules. Thus, if not true self-replication, the system could still be dynamic enough to go on with its mission for much longer by replacing damaged parts and handling unforeseen circumstances. These could also be beneficial here at home, doing useful work in remote areas, deep underground, and so on, but are likely most useful where humans cannot be easily used instead and where no support infrastructure exists. Some imagine clanking replicators as doomsday machines, perhaps getting out of control, killing their masters or used maliciously by them, and then going off to turn significant amounts of usable energy and matter in the universe into more copies. Some even go so far as to consider this as one answer to the Fermi paradox. The Fermi paradox wonders why we haven't been contacted by aliens yet or detected their presence; presumably, the whole galaxy should be full of them, given that even at sub light speed, an advanced race could establish colonies all over the milky way in well under a million years. One conjectured answer is that alien races hide themselves to avoid attracting the attention of clanking replicators, possibly sent out by other, hostile, aliens. This highly speculative notion assumes much, but it's very interesting indeed. Greg Bear wrote an excellent two part sci-fi series about just such an idea, “The Forge of God” and “Anvil Of The Stars”. It makes me think twice about announcing ourselves to the universe, as we now do, both accidentally and on purpose... In the next part, I'll talk about artificial life of the kind most people think of – the kind that's more in line with what we now consider 'alive': biological life. -- The WikiPedia page about clanking replicators. This is a good place to start if interested in the subject: http://en.wikipedia.org/wiki/Clanking_Replicator This is one of the links found on the WikiPedia page: "3D printer to churn out copies of itself": http://www.newscientist.com/article.ns?id=dn7165 ------------ About the author Frederick Smith: I enjoy writing about the positive virtues of humanism - humanists are the good guys. This is my second foray into the UK writing discordia. This time around, I want to be a tad more raw - maybe a bit edgier (does that sound "art-see"?) Maybe I'll address even more issues that most Americans consider taboo... About my personal background and life: I was born, I got some education, worked, ate, and had some kids. It seems I like to write – something that was unknown to me until relatively recently...How's that for detail? ;) Like so many these days, I too have a blog! But, I haven't updated it since the day after I signed up for it, so I won't bother to give out the link. Hate mail is welcome unless you are from the Army Of God. Please! It's not that I mind seeing pictures of aborted fetuses in my inbox, but once you've seen one you've pretty much seen them all... Email: dahlek65@yahoo.com Tell a friend about this site! ------------ All articles are EXCLUSIVE to Useless-Knowledge.com. Please link to this article rather than copying and pasting it onto your site (which would be unauthorized and illegal). |
||||||
|
|
|||||||
|
