Walking through New York City during his years as a postdoc, Albert-László Barabási was intrigued not by the glitz and glamour, but by the notion of the invisible networks of cables and pipes necessary for modern life. Given that he trained as an engineer in Bucharest, Romania, such thoughts were not unusual for the Hungarian native. (see CV)

But it is the connection between networks and another passion of his — chaos and fractals — that really underpins his scientific career. His fascination for fractals led him to Budapest to work with international fractal expert, Tamás Vicsek. Later, in New York, he married the ideas of the self-replicating patterns in fractals with the structure of networks.

Left to his own devices at IBM, Barabási began asking himself: “What the heck is a computer?” Realizing that everything from computers to electricity distribution to water pipelines is networked, he wondered why nobody in science had paid much attention to networks. “Networks must not be random, but we didn't know anything about them,” he says. So he decided to fill in the gaps.

Timing was on his side. The appearance online of digital maps of the Internet and biological networks in the late 1990s became the foundations of his work. The result was two papers introducing the concept of scale-free networks. No matter what system he looked at, Barabási discovered that all networks are dominated by a few highly connected nodes or hubs.

Barabási decided to devote his full attention to networks, even though he lacked both funding and tenure. He credits this bold move with his ultimate success. “I thought this is going to be more important than anything I've done before, I can't do it half-hearted or half-brained,” he says.

Now Barabási has his eyes on the next challenge: a theory of complexity. “I believe that if there will be a theory of complexity it will emerge in the next ten years,” he says. “How do I position myself to contribute to that goal?”

He is optimistic that his one-year move to the Dana-Farber Cancer Institute at Harvard University, will help him. As the field has advanced, Barabási says, it has become increasingly important to be close to experimental groups generating data.

For those who want to follow in his footsteps, Barabási offers one piece of advice: aim very high. “You'll never reach that very high aim, but if you reach 75%, it's still very good,” he says. “If you aim low, 75% gets you nowhere.”