Affiliation(s): Microsoft Research
Abstract:
The movies, it turns out, are a terrible model for data visualization.
Let me step back a moment. If I want to know what the future of, say, intelligent agents might look, I have a lot of choices. KITT from Knight Rider, or the Star Trek computer, or HAL from 2001, or any of a thousand other films and television shows will all give me examples of how speech recognition and intelligent agents might look. A designer of current systems can push back, or pick points on the spectrum—“I’d think it can be more mechanical, less humanoid.”
What about computer graphics? The Holodeck. R2D2 projecting the Princess Leia. Infinite zooming in Blade Runner. 3D worlds in Jurassic Park and a million other movies.
And so it goes for lots of developing technologies. Flying cars and self-driving cars. Robots and tablet computers. Movies have shown us visions of the future for power plants, and long distance transport, and food preparation–and even for how doors might work. Film directors, screenwriters, and effects teams have done a wonderful job of portraying the a computerized, high-technology future.
Now, since the beginning, we’ve all understood that computers are very good at presenting and storing information- Or, at least, we’ve believed that we understand that. Sadly, we have only the poorest of examples to work from.
Continue reading