31 July 2000
Reported, written and edited by David Duberman
for editorial/subscription inquiries, send mailto:email@example.com
Search the Spectrum archives at http://www.3dlinks.com/spectrum
SIGGRAPH (hereafter: Siggraph: http://www.siggraph.org) is the Special Interest Group for Graphics of the venerable professional organization Association for Computing Machinery. It's also the name of the group's annual trade show for computer-graphics (CG) fans of all persuasions, from grizzled veteran hackers who cut their eyeteeth on computer punch cards to starry-eyed students. But Siggraph isn't just about CG: It also concerns itself extensively with computer-human interaction. This point was resoundingly brought home by Raymond Kurzweil, the keynote speaker at this year's event, held last week in New Orleans.
Kurzweil, an inventor, artificial-intelligence expert, and all-around visionary, titled his talk "The Human-Machine Merger: Why We Will Spend Most of Our Time in Virtual Reality in the 21st Century." He predicted a day, not extremely far in the future, when computers and humans will merge seamlessly. You might think this sounds rather far-fetched, but Kurzweil made a convincing case by presenting statistics that seemed to indicate that the pace of technical change has been accelerating, doubling the rate of progress every decade, and will continue to do so.
The inevitable result of the ongoing trend toward miniaturization will result in nanotechnology, he says, and we'll probably have self-replicating nanotechnologies in 25 years, not 100, as some predict. But before that, possibly in less than 10 years from now, we'll have tactile-based virtual reality. In discussing VR, he made the point that, despite those who claim that VR participants act irresponsibly, the technology is analogous to the telephone, and that people who make verbal commitments by phone generally meet them.
According to Kurzweil, the next major advance in digital technology will be three-dimensional computing, which works like the brain, claiming he's seen prototypes of such machines with 300 layers of circuitry. He predicted that a one-inch nanocube, potentially feasible by the year 2030, would be 1,000 times more powerful than the human brain. He then went on to talk about brain scanning via nanotech, which will ultimately allow us to reverse-engineer the brain, creating/re-creating the types of processes that actually occur in our most important organ. Thus, machine intelligence won't be artificial, but derived from human intelligence, and will have a natural, human feel. (Are you there, Hal?)
From these advances we'll have access to virtual reality that connects directly to our brains, courtesy of nanobot implants. This will encompass all of our senses, and even extend to emotions. Of course, the highly optimistic talk didn't touch on how a futuristic Hitler could capitalize on such access to the brains of millions or billions, but no doubt security will be foolproof by then (ahem).
The talk, based (I think) on Kurzweil's recent book, The Age of Spiritual Machines: When Computers Exceed Human Intelligence, was well received by the 25,000+ show attendees, especially when he concluded by making a firm connection between his vision and his audience. His final predictions were that VR would be a graphics challenge, and that graphics would be 50 percent of computing.
The topic of virtual reality was also covered on a more mundane level at Siggraph in the form of various panels and presentations. A panel titled The Actual Reality of Virtual Reality Software was moderated by the irrepressible Linda Jacobsen of SGI, who likened VR software to Rodney Dangerfield: both have a cult following, neither became stars until middle age, and neither gets any respect. Panelist Ken Pimentel of Engineering Animation, Inc. bemoaned VR's lack of user-interface standards, concluding that only application experts can derive the value of the visualization experience, and that these experts can't use their knowledge on the next immersive application. Kent Watsen of the Naval Undergraduate School talked about general-purpose toolkits for creating virtual environments (VE), saying that "Because these toolkits have different architectural implementations, they have fragmented the VE community." On a more optimistic note, he predicted that "component-oriented programming, a recent trend in software engineering focused on establishing the 'standards for interoperability,' may be the proverbial 'silver bullet' (Brooks, 1987) the VE community so desperately needs."
Of course, once we've built our virtual realities, there remains the problem of what to do with them. Many believe that the ability to participate in interactive storytelling holds the answer. A special session titled Fiction 2001, somewhat overpopulated with nine panelists, held the sizable audience's attention with imaginings of the future of fiction in the emerging realms of networked computers. Novelist/actor/director (and Microsoft researcher) Andrew Glassner supposed that interactive fiction is a problem, because, while most stories are about conflict, most people are conflict averse. What's more, he said, improvisation is very difficult. He offers the theater piece "Tony and Tina's Wedding as an example of the ultimate form of interactive storytelling because audience members can say anything they want to the actors, and real, trained people respond.
Also in Fiction 2001, author Espen Aarseth made the case for games as story. Jesse Schell of Walt Disney Imagineering reinforced this notion with his emphatic belief that "The videogame is the most exciting medium of our time." AI researcher Phoebe Sengers (sp?) predicted a battle of ideas between corporate retellers of "tired twentieth-century narratives" and insurgent writers and artists who will create "atomized, decentralized, emergent post-narratives, elegant in construction, intellectually dense, extremely hip, but not exactly a pleasure to read." Her hopes are for a third alternative, "technology that supports our need and desire to tell [meaningful stories]." The final word on the topic came from Jay David Bolter, who stated, "We cannot know the future of fiction, but we can talk about its present condition. What strikes me is the number and variety of narrative media forms today: from traditional novels and films to websites and interactive games. There is no one form around which our culture's creative energies are converging." You can find more from this and several other Siggraph panels at http://www.mrl.nyu.edu/noah/s2000/. Incidentally, this site uses a Java applet for a clever zooming effect that, alas, doesn't seem to work with Netscape.
Siggraph attendees had a chance to participate in interactive storytelling via a fascinating experiment called Terminal Time, created by three researchers whose names I unfortunately neglected to record. In this interactive cinema piece, the computer poses a series of psychographic/demographic multiple-choice questions to the audience, whose members applaud for the chosen answer. The choices that receive the loudest response "win," and thus drive the succeeding cinematic segment. The resultant short movie is a "world" history of sorts, selected from over four hours of digital video, accompanied by a synthetic-voiced narrative. Both are pieced together by an artificial intelligence program based on the audience responses to questions like "Things always work out well in the end (agree, disagree?)."
The creators ran the audience through the exercise twice so we could choose different answers each time, and see how the program changed. Indeed, the results differed vastly between the two sessions. Unfortunately, the synthetic-voice narration, although it was the best the creators could find, was often difficult to understand, forcing the audience to work hard to hear it. Still, it was fairly obvious in many cases how the software slanted the events and sequences of history based on the audience responses, driving home the oft-observed thought that history is never objective. An unexpected side effect was that noting your fellow audience members' choices was often more interesting than the video. Also, it was easy to skew results by clapping loudly, with the result that sometimes, less-popular choices "won" the vote.
It came a day before the actual end of the show, but the session "Phil Tippett's History of Animation" closed Siggraph with a bang for many attendees. Tippett, a veteran creator of cinematic special effects who's worked on epics from Star Wars to Starship Troopers (and beyond), is now the proprietor of Tippett Studio, a special-effects production house in Berkeley, and dreams of someday producing full-length "fake … er, digital" movies. In a mobbed hall, he presented a sequence of film clips designed to give perspective to animators who may have little idea of what's come before. As Tippett showed the clips, he often paused to offer ideas or reminisce about his experiences in the biz. He's a refreshingly candid speaker, and addressed the audience as if he were chatting with an old friend across the dinner table.
Tippett started out with the generally acknowledged father of film animation, George Melies, and then proceeded to show clips from such varied artists as Winsor McKay, Latislov Sterovich (sp?) of Russia, Willis O'Brien (best known for King Kong), the Fleischer brothers, George Pal, and, of course, Ray Harryhausen. A special treat for the audience was Tippett's own tribute to Harryhausen, made for the latter's 80th birthday party, which showed the skeleton warriors from Jason and the Argonauts singing "Happy Birthday." From his own oeuvre, he showed excerpts from Star Wars, Robocop, Dragonheart, and others. At the request of an audience member during the Q&A, Tippett promised to post a list of his animation history recommendations on his Website (http://www.tippett.com/), but it wasn't there yet at the time of this writing.
The high point of any Siggraph is the Electronic Theater, and this year's was no exception. The films, showing a wide range of CG applications, from entertainment to science and engineering, are generally available for viewing later on in a roadshow, so I won't spoil any of the surprises for you. But be sure not to miss "Lucie," with the most realistic character modeling to date, and the hilarious "For the Birds," from the CG humormeisters at Pixar.
In part two of this special Siggraph report, coming later this week, we'll provide a description of some of the entries in the Emerging Technologies program, plus a listing of the various new/updated products and announcements at the exhibition.
Spectrum is an independent news service published every Monday for the interactive media professional community by Motion Blur Media. Spectrum covers the tools and technologies used to create interactive multimedia applications and infrastructure for business, education, and entertainment; and the interactive media industry scene. We love to receive interactive media and online development tools and CD-ROMs for review.
Send your interactive multimedia business, product, people, event, or technology news to: firstname.lastname@example.org. We prefer to receive news by email but if you must, telephone breaking news to 510-549-2894. Send review product and press kits by mail to David Duberman, 2233 Jefferson Ave., Berkeley, CA 94703.
If you contact companies or organizations mentioned here, please tell them you saw the news in Spectrum. Thanks.
Please send address changes (with old and new addresses), subscribe and unsubscribe requests etc. to the above address. If you use the Reply function, please do _not_ echo an entire issue of Spectrum with your message.
Publisher's note: We are now accepting limited advertising. If you'd like to offer your company's products or services to Spectrum's elite audience of Internet and multimedia professionals, send an email query to mailto:email@example.com, or telephone 510-549-2894 during West Coast business hours.
- David Duberman
©Copyright 2000 Motion Blur Media. All rights reserved. No reproduction in any for-profit or revenue-generating venue in any form without written permission from the publisher.