Last month, DARPA hosted Wait, What? A Future Technology Forum in St. Louis. There, 1400 people gathered for a national discussion and showcase of new ideas and advances in the technoscape, among them optical techniques for seeing around corners and neural interfaces that allow people with paralysis to control a prosthetic limb by thought alone. At the forum, DARPA’s Defense Sciences Office (DSO) ran a pair of breakout sessions titled Science, Disrupted: Beyond the Limits of Intuition, Computation and Measurement, with the hope of learning from attendees what they imagine could throttle up science into an even more powerful engine of discovery and technology than it is now.
Some of the input from participants — as well as from others responding to survey questions posted on the meeting’s app — aligned with major themes of the meeting, most notably the emergence of artificial intelligence as a necessary assistance for human beings to understand phenomena as complex as genomics, their own brains, and the coming Internet of Things in which even chairs and toasters will have their own IP addresses. Other responses were even further out of the box, which is territory that appeals to DSO program managers and the rest of the DARPA community.
“The correct, consistent theory of quantum gravity will be worked out and the origin of our universe will be resolved,” remarked a university physics professor. He was referring to the quest to finally uncover a quantum-scale basis for the gravitational force that operates even on the largest objects and structures of the universe. “This quantum gravity theory will likely predict the existence of other universes as well, that is — a multiverse,” he mused. For those who don’t think the universe as we know it is enough, this surely was a heartening prediction. On the other hand, if there happens to be adversaries in other universes and we find portals to them, the threat space with which DARPA will need to concern itself could inflate to unprecedented dimensions.
Focusing instead on the inner universe of humanity, the director of a startup consulting firm predicted the impacts of what she called “direct brain interfaces,” which stand to connect our brains to other interfaces. “Linking even more closely through our technologies with each other, with the [Internet of Things], even with other species, will vastly change perceptions of self,” this participant wrote. “Since the definition of self and in-group is central to motivation, resource allocation and actions will be driven by different values.” Several respondents pointed to perhaps the most extreme possible disruption to humanity’s definition of self: the discovery of life elsewhere in the universe. Such a revelation, they variously opined, would “change everything” and “revolutionize the human understanding of the origin of life.”
In a similar vein, another independent consultant at the meeting observed that two red-hot fields, computation and neurobiology, are merging into what he said could be one of the greatest disruptions of the near future. “We already have crude capabilities to print circuits on skin and organs,” he stated. “We will increasingly be able to interface computation to, and make sense of, human cognitive, emotional and physiological processes. The power this provides will unleash massive changes in how and what we compute, and in both human-machine and human-to-human communication. It will also create powerful and potentially ethically-challenging monitoring capabilities.”
One federal scientist opted to focus on life as we actually know it and to which we have direct access here on Earth. “The conceptual and computational horsepower to achieve first principles modeling and engineering of biological processes, from the large-molecule level on up” will constitute a huge disruption that will have enormous practical consequences, this attendee predicted. The necessary combination of “ingredients” for this disruption—“an explosion in computational hardware, novel algorithm development, and — perhaps most important — lots of profit motive in the form of potentially huge medical advances” — are becoming available, he argued.
Some participants at the breakout sessions suggested that even seemingly simple changes in scientists’ behaviors could go a long way toward revving up discovery. A materials chemist, for example, suggested that a genuine community value of sharing raw data — not just the interpreted data published in scientific papers — would greatly expand opportunities for computers to apply machine learning to identify new technologically-promising molecules and materials.
“Chemists are great with intuition,” another attendee shared and then mulled: “Can we teach machines the chemists’ algorithm?” That catalyzed yet another engaged participant to suggest that for any machine to become a good chemical “thinker” it will have to incorporate randomness (stochasticity) into its algorithms as a proxy for the still-mysterious intuitive leaps human chemists so often pull off when they design new reactions or pose the possibility of never-before-seen molecular structures. “We come up with solutions in our sleep and in the shower,” quipped DSO program manager Jim Gimlett, implying that computers might become discovery engines if they had machine equivalents to those sorts of intuitive leaps ... or perhaps to sleep and showers.
As far out as many of these ideas might be, it would be unDARPA-like to dismiss them. As one Wait, What? attendee put it in the forum’s smartphone-accessed activity feed, “Everything is Sci Fi until it happens.”