OT: Machines as smart as monkeys, or men - "people will soon learn to love them"greenspun.com : LUSENET : TimeBomb 2000 (Y2000) : One Thread
(EDUCATIONAL USE ONLY: Copyright Guardian Media Group, plc)
LIFE WITH THE CREATURE FEATURE: Now they're cute and cuddly and live inside your computer. But, writes Jack Schofield, the rise of the robots could have serious consequences for us all.
("The Guardian", Thursday January 13, 2000)
The creation of artificial life will be one of the great themes of the 21st century. It is already helping to drive the market for toys and games, from CyberLife's Creatures to Tamagotchis to Sony's AIBO robot dog, while "intelligent agents" (software robots) are expected to become indispensable on the internet. Indeed, a generation of children is already becoming acclimatised to A-life through keeping "virtual pets" who "live" in personal computers. Many keep dogs and cats as mess-free screen pets, thanks to Mindscape's Dogz and Catz programs, while Creatures allows them to explore breeding and genetic manipulation using Norns.
Steve Grand, the British designer of the Creatures program, says: "Probably some of the best artificial life scientists in the world are 13 year old American girls, because of what they've done with Creatures. There's even a Norn genome project like the human genome project, and some people are doing mutation tracking."
Toys also allow companies to experiment with advanced technologies that may not be ready for hard commercial use, such as speech recognition. For example, you can talk to Fujitsu's Fin Fin, a parrot-like "virtual pet" with a head like a dolphin, and Lernout & Hauspie's Talking Max2, a cartoon parrot, even talks back. Max needs to be trained, of course, but he can obey verbal commands and engage in witty conversations. (Both the Fin Fin and Max2 programs come complete with microphones.)
As computers become more powerful, designers will be able to simulate more sophisticated life forms, including people. Just before Christmas, Mindscape launched Babyz as the natural progression from Dogz and Catz. Other companies, such as Dave Morris's AIdeaLabs in Texas, have already tried "virtual girlfriends", and Electronic Arts is about to launch a people simulation called The Sims, which has been developed by Will Wright, founder of Maxis and creator of the Sim City series.
As David Crane discovered when he pioneered the idea of keeping a person as a pet in Little Computer People - launched in the 1980s for the Commodore 64 home computer - the problem is making the simulation varied and believable enough to keep the player's interest. And the way to do that is to give the creatures the ability to learn, and to reward them for learning things, which is how Creatures works.
When they become smarter, of course, artificial life constructs will have to go out to work. They will probably become a standard part of computer interfaces, and make it easier to do things on the internet. Microsoft's Office suite and Word 2000 already provide an (admittedly stupid) "intelligent agent" as an optional interface to the help system, and users can substitute a cartoon dog, or a companionable cat for the now-familiar animated paperclip. The same Microsoft Agent technology is also being used experimentally on dozens of web sites.
There is a huge potential market for such AI (artificial intelligence) constructs. Software robots - bots for short - will provide intelligent opponents for games players, especially in multi-user online games. They'll scour the web for information, like the ones pioneered by another British company, Autonomy. They'll go comparison shopping to find their owners the best prices. Eventually they'll become smart enough to answer the phone and much of your electronic mail.
But some people are already worried about what happens when they become too smart. Professor Hugo de Garis, a "brain builder" based in Japan, believes that what will actually dominate the century is a battle for "species dominance". He says: "Humanity is going to have to decide whether it does or doesn't build artilects (artificial intellects), which are coming simply due to the rise of the technologies. It's the Terminator scenario . . ."
Will bots and robots still be our pets, like the Robo Kitten that de Garis is developing? Or will we be theirs? Will people amuse themselves by breeding and educating artificial life forms like the Norns in Creatures, or will they amuse themselves by breeding and educating us?
Such worries sound silly. After all, AI research has been going on for decades without making an impact outside a few specialised areas, such as oil exploration and computer chess. Why should that change now?
"Computer power," replies professor Hans Moravec, director of the Robotics Institute at Carnegie Mellon University in Pittsburgh. "I think the primary bottleneck was computer power. It's a bit like having a rubber band engine to power a 747: it doesn't matter how good your aerodynamics is!"
Moravec argues that generally AI has not had access to more computer power since the 1960s, when research was done using $10m supercomputers and mainframes. Research moved to minicomputers, workstations and then microcomputers, each of which was dramatically cheaper but not inherently more powerful. In the 1990s, however, micro- computers increased in power from less than one MIPS (the ability to perform a million instructions per second) to 100 MIPS, and they're now approaching 1,000 MIPS.
In fact, computer power is increasing so rapidly that Moravec is turning down speaking engagements and other distractions to concentrate on developing navigation systems for things like guard robots and cleaning machines.
"I'm rushing to capitalise on our decades of preparation before further increases in computer power make the problem so easy everyone will be doing it!" Moravec says.
"The business plan says that with 1000 MIPS we can build that kind of thing. We'll have 1000 MIPS microcontrollers in about three years, and we'll go commercial at that point."
Moravec typically compares the computer power available for AI and robotics with animal life forms. The 1 MIPS machine - think of a 1985 personal computer - is about as smart as a nematode worm. The new 1000 MIPS systems have advanced to insect level, perhaps approaching a spider. Moravec thinks he'll be able to apply the power of a guppy - "not as smart as a goldfish" - to building something like a robot vacuum cleaner in about 2005.
The comparisons highlight the rate of progress. Computers are making the kinds of advances in five or 10 years that took evolution 25 or 50 million years.
After another one or two thousand-fold increases, it's possible to imagine machines as smart as monkeys, or men. Moravec thinks we'll reach that stage in about 2040.
"I strongly believe that the development [of human-level robots] can be done incrementally," says Moravec. "That was how intelligence was built the first time, by evolution. So what I look for is a series of slightly more sophisticated machines, each capable of earning a living."
Steve Grand - who, at 41, was given an OBE for Creatures in the New Year honours list - has also decided to go into robotics, and this week announced a new company called CyberLife Research.
"I want to make machines that think," says Grand. "I don't mind doing it in software, but I want to do it in hardware as well. I'm trying to develop a general-purpose brain on a chip. I'm not looking for applications right now, but I'm thinking of building a model glider that can learn to fly - hopefully before it hits the ground."
Grand is sceptical about the field he's just entered - "it took evolution millions of years for good reasons" - but he agrees that 40 years is a reasonable time-scale for brain building. "We should be at least at chimpanzee level."
Hugo de Garis thinks he's leading the way, having formally switched on the CBM or CAM-Brain Machine (CAM stands for Cellular Automata Machine) on November 29 last year. This has been developed at the Evolutionary Systems Department of ATR Labs in Kyoto, Japan, and imple mented by Dr Michael Korkin's Genobyte Inc in Boulder, Colorado.
The idea, in essence, is to build machines that create brain-like structures called neural networks by generating them automatically. De Garis says the CBM "can evolve a neural net circuit module of some 1,000 neurons in a second or so". The aim was to have an artificial brain with a billion neurons by 2001, though 75 million neurons now looks the likely maximum.
Building the brain is only the first stage. "Evolutionary engineers" then need to turn the raw material into a working brain, before programmers can use it to control something like the Robo Kitten. All the kitten's behaviour patterns need to be specified, for example. De Garis says this should be running to provide "proof of concept" in two years.
De Garis sees the scope for perhaps half a dozen brain-building projects in different countries, though the next two CBM installations will both be in Belgium. One will be at the Starlab in Brussels, where de Garis will become chief scientist at the end of this month. The other has been bought by Lernout & Hauspie, which specialises in speech technology, and which includes Intel and Microsoft among its investors. But to create really useful artificial brains, de Garis thinks we need "national brain-building projects - something equivalent to Nasa".
As well as being a brain builder, de Garis is also raising the alarm about the dangers of brain building: "I'm very much part of the problem," he admits.
"In the short term, the products will have very beneficial effects: they'll be just wonderful," says de Garis. "And the machines will get smarter and smarter year by year, until eventually they will become so smart they take over their own development. At that point, humans lose control. I think and think about how this terrible scenario can be avoided, but I just can't see a way out of it."
Steve Grand is more optimistic. "They're just another life form, and there are millions of other species out there who haven't 'taken over the world'. People are bound to be resistant at first, but then they're also scared of impersonal technology, and one of the good things A-life can do is make machines more personable and friendly. People will soon learn to love them!"
-- Risteard Mac Thomais (firstname.lastname@example.org), January 14, 2000
"...And the machines will get smarter and smarter year by year, until eventually they will become so smart they take over their own development. At that point, humans lose control. I think and think about how this terrible scenario can be avoided, but I just can't see a way out of it."
...and this is why God invented what we call "The On/Off Switch," with the backup system known as "The Electric Plug." (Although the idea of a virtual boy/girlfriend with the intelligence of a chimpanzee does bring some interesting images to mind.)
-- Colossus/Guardian (We@Are.One), January 14, 2000.
What is the Matrix?
Wish I had taken the OTHER pill, heh.
-- Hokie (Hokie_@hotmail.com), January 14, 2000.
There once was a group of proud scientists. One day they were all sitting around recounting the degrees and the awards they had garnered, and they came to the realization of just how intelligent man had become. It ocurred to them that man, through science, had achieved a level of development where God was no longer relevant or needed. They discussed the matter further, ultimately deciding that it was time to send a scientific delegation to God to suggest his "retirement." So the scientist says to God : "God, you know, a bunch of us have been thinking and I've come to tell you that we really don't need you anymore. I mean, we've been coming up with great theories and ideas, we've cloned sheep and monkeys, and we're on the verge of cloning humans. We are even creating artificial intelligence, which, some day, will surpass its creator in smarts. So as you can see, we really don't need you anymore." God nods understandingly and says: " I see. Well, no hard feelings. But before you go let's have a contest. What d'ya think?" The scientist says: "Sure I'm all for it. What kind of contest?" God: " A man-making contest." The scientist: "Sure! No problem" The scientist bends down and picks up a handful of dirt and says: "Okay, I'm ready!" And God says: "Oh, no. You must start with your own dirt."
-- anon (email@example.com), January 14, 2000.