HALF MOON BAY, Calif. As biology and electronic engineering start to converge in the nanotechnology world, a "real problem" is emerging because the two methods of developing systems are essentially incompatible, a leading venture capital said Monday (Jan. 8).
Steve Jurvetson, managing director of Draper, Fisher Jurvetson, said the trend comes because traditional electronic systems and even many nanotech systems are designed from the top down. In biological systems, it's an evolutionary methodology.
"There is a real problem here," he told the annual Industry Strategy Symposium (ISS) here. "A lot of time in the engineering world we like to design things and have them do we want. The concept of evolved systems where there's a bit of an out of control element is a bit distinct" from that.
Jurvetson pointed out that an evolved system, like a neural network gets really complex really fast and can be very difficult to understand. "You can't understand psychology but studying a neuron," he said.
A simulation of an evolved system may be as complex as evolving the system it itself, he said. "You may just have to get to the point of being comfortable with the evolved system, which is spooky," he told the audience. Jurvetson published a paper on the topic last summer in the MIT Technology Review.
A "bio-nano" future will involve a blending of information technology and biology in which an interdisciplinary "learning-doing" cycle will emerge. He also suggested an entirely new way of thinking is demanded because inventors will not evolve or grow complex systems with the kind of huge code base used in Microsoft office.
That said, the Moore's Law-type ramp lends itself nicely to biological systems, Jurvetson noted. The cost to synthesize base DNA pairs has fell from $12 to under $2 between 2000 and 2005, he said. Scientists "can do the equivalent of a human genome every three months," Jurvetson said. "Moore's law is driving this field as well."