Dan Reed heads a crew within Microsoft that may have the coolest name of any division in the company: the eXtreme Computing Group, or XCG. Formed a little less than two years ago, the group is part of Microsoft Research, but its mandate goes beyond R&D: it’s to help the company as a whole look into the future and question its assumptions about the nature of computing.
Which makes Reed the perfect panelist for Beyond Mobile: Computing in 2021, Xconomy San Francisco’s marquee information technology event this spring. Set for next Tuesday, May 17, on the campus of SRI International in Menlo Park, this evening forum will give audience members a chance to interact with Reed and a cast of brilliant thinkers, researchers, and entrepreneurs on a question we’re all usually too busy to think about: just where is all this technology taking us? Given the current pace of progress in semiconductor and software technology, what sorts of capabilities will computers have 10 years from now? And what can we do now to plan for the entrepreneurial opportunities—and social and political challenges—that these capabilities will create?
Reed—a former UNC Chapel Hill professor and a onetime member of the President’s Council of Advisors on Science and Technology—is uniquely qualified to talk about those questions, because he’s not just a technologist. He’s also Microsoft’s corporate vice president of technology strategy and policy, meaning he spends a lot of his time thinking and talking about how trends in computing technology translate into issues that politicians and businesspeople are going to have to work through at some point.
I spoke on the phone with Reed last week, getting his views on everything from smart radios to data center design to AI. We’ll delve even deeper into many of these issues at next week’s event, so be sure to get your ticket now. Meanwhile, as a preview, here’s an edited summary of my talk with Reed.
Xconomy: Being a futurist is notoriously hazardous. Bill Gates himself said, “We always overestimate the change that will occur in the next two years and underestimate the change that will occur in the next ten.” What do you do in your own work to try to work around all the uncertainties and biases involved in technology forecasting?
Dan Reed: One of the things I do is look at exponentials that have been in flight for a long time and ask how long they can continue. I think our semiconductor roadmaps are an example of that, looking at transistor size and some of the quantum-mechanical issues that are starting to rear their heads, and the power issues. When people start to struggle with something, then there is a chance that a phase transition is about to occur. It doesn’t mean that some breakthrough won’t occur to obviate the conventional approach. But when you see the bag of tricks starting to look empty, then I ask questions. It’s time to point out that this field may be in for a disruption.
I also talk to people outside technology. I talk to my friends in the arts. What do they find interesting? I try to listen to users of technology who aren’t technologically sophisticated to see what unexpected things they’re doing. And I do a lot of general-interest reading and trying to understand societal trends, and looking at where some of our social structures are being challenged.
X: What benefit do you get from talking to people outside of technology or engineering?
DR: I think sometimes people outside a domain are actually better at predicting than people inside it. They’re usually less technologically knowledgeable, therefore they can make wildly inaccurate assumptions. But at the same time, they are not encumbered by the conventional wisdom that those of us in the discipline may have.
X: How would you describe the mission of the eXtreme Computing Group?
DR: Craig Mundie [Microsoft’s chief research and strategy officer] and Rick Rashid [senior vice president of research] originally asked me to come to Microsoft to take a blank sheet of paper on how to build a data center. They said, “Don’t take any of the
