Less than a year after he got his high school diploma and left Shenandoah Junction, W.Va., for Silicon Valley, Robbie Barrat began teaching computers to paint. He fed a few thousand examples of paintings into his artificial intelligence software until it learned how to create landscapes like the one on this issue’s cover. By computer standards, these works of art took a long time to produce: a little more than two weeks. “AI is going to be one of the larger art movements of this century,” says Barrat, a Stanford researcher who goes by @DrBeef_ on Twitter. “It just has really great untapped potential.”
This is the world AI is making. In this year’s Sooner Than You Think issue, we focus on the field’s dreamers, die-hards, and driving forces, from its early architects in Canada (and a forgotten one in Switzerland) to the latest experiments by the world’s biggest companies. We look at the Chinese chips, Israeli software, and British aerodynamics shaping the frighteningly smart machines of tomorrow. We also explore the field’s ethical dilemmas, which in those circles are too often mere afterthoughts. And we try to make Google Calendar predict our next moves based on our past moves. That was humbling.
Early AI experiments show machines have their own share of moments that seem almost as grim as a careful review of your day planner. Still, Barrat’s software left him hopeful for a new era of human-AI collaboration. “It would go through these bouts of generating really colorful paintings and really dark, gloomy paintings,” he says. “It eventually got out of that.” If computers can fight through a blue period, there may be hope for us, too.
Leave a Reply