Saturday, January 28, 2006

Errant SF Futurism and Our Dumb Computers: Notes Toward a Review of Accelerando

by Tom Bozzo

Near-future futurism is risky, in that some or most of the audience will eventually get to see what turned out, what was reasonable to expect but didn't turn out, what was (for one reason or another) totally full of crap, and what else was totally overlooked.

For example, pop in the DVD of Blade Runner (1982), famous for its visual design of Los Angeles in 2019 featuring flying police cars and a forest of futuristic skycrapers dominated by the enormous pyramidal headquarters of Tyrell Corporation. Even though it's a bit early to make the call, the distinctive visual elements can be classified as totally full of crap. What may seem to be bizarre feats of bio-engineering can't be ruled out, though it's doubtful they'll take the form of creating genetically-engineered people lacking the usual civil rights. Nuclear apocalypse? (*) We can keep our fingers crossed.

Almost totally overlooked was the Moore's Law implication that computer chips will sport billions of transistors (and soon) — our ubiquitous microelectronics, not to mention macroelectronics like giant flat-screen TVs, are conspicuously absent. The scene where Rick Deckard uses voice commands to navigate a 3-D snapshot on some sort of computing device is almost painful given knowledge of modern 3-D computer games.

Obviously, some artificial intelligence stories, notably 2001: A Space Odyssey, have suffered not just because we have had difficulty getting past the tin cans in low-Earth orbit stage of human space exploration (**), but because our now ridiculously powerful computers (by the standards of any time sufficiently far in the past, say 10 years or more) remain really stupid.

There are plenty of silly things that happen to be highly computation-intensive. For instance, I created this image in support of a possible dorktacular future post; it took 2 hours, 6 minutes, 55 seconds to render on my 2GHz iMac G5. The iMac has variable-speed cooling fans that provide some indication of how hard the computer is working at any given time — not very at all, in nearly every use to which it's put, but they were running at full tilt for those two hours. This places part of the blame on software designers, who could be said to have failed to come up with good uses for zillions of idle processor cycles here and on nearly every other desktop.

Which takes us to Charlie Stross's starting point in Accelerando, the inaugural Marginal Utility book club reading material. (***) It's a world of a few years hence that closely resembles our own, but with better computers and ubiquitous very high-speed wireless internet access. That, of course, is perfectly plausible. Indeed, our wireline telephony oppressors seem hell-bent on pushing us into that world.

From there, Stross takes the notion of 'accelerating returns' a la Kurzweil "seriously" — in the sense of creating a fictional world in which it should be taken seriously because it describes the actual pace of technological progress — and things get weird fast. Much of the weirdness stems from machines' ability to extend and even subsume human consciousness. (****) For that to happen, there needs to be an enormous qualitative leap in which the transistor population boom (and the corresponding computation boom) moves us past the ability to add lots of numbers very quickly and the relatively simple (if useful) 'bots the current computational technology allows.

Since the technological transition in Accelerando happens roughly 10-15 years from now, one question is, what will anyone make of the novel in 25 years? Will it be regarded as an advanced version of pre-rocketry accounts of trips to the moon?

-----------------------------

(*) The Philip K. Dick source material, Do Androids Dream of Electric Sheep? (1968) made it clearer than the film that the setting is postapocalyptic.

(**) Not without good reason, as robots don't need to worry about being kept alive in environments that are extremely hostile to terrestrial life, and notwithstanding John Tierney's column in today's Times (shorter Tierney for non-Times Selecters: Real science is boring, so what we really need are a few rich people to ignore the lesson of Bill and Melinda Gates and waste enormous sums on a great big adventure).

(***) Pub Sociology's Brayden King joined in but reports not having completed the novel. Since Accelerando consists of four works previously published as novellas (or long short stories) plus some additional connective prose, portions can stand alone to a greater-than-usual extent.

(****) Stross, who maintains two blogs, posted a spoiler here.
Comments: Post a Comment

<< Home

This page is powered by Blogger. Isn't yours?