Pixel Qi Out to Bring Principles of Inexpensive Laptop Design to Consumer Market: Former One Laptop CTO Mary Lou Jepsen On Her New Startup
If only laptops could run on qi—the spiritual energy that, in traditional Chinese philosophy, pervades all things.
Well, if anyone has come close to making that happen, it’s Mary Lou Jepsen, founding chief technology officer at the Cambridge, MA-based One Laptop Per Child Foundation (OLPC). At the foundation, Jepsen did what computer-industry executives said couldn’t be done: design a powerful laptop for children in the developing world that worked on less battery power, and for less money, than anything the major PC manufacturers could create. Last November, the foundation’s neon-green, rabbit-eared XO laptop went into mass production in China. And on New Year’s Day, Jepsen’s long-planned departure from OLPC became official.
But she hasn’t gone far: the next gig for Jepsen, former CTO of Intel’s display division, is Pixel Qi, a Hull, MA-based startup she has created to design and build components for low-cost information devices that could be sold to consumers right here in the United States, as well as to people in the developing world. Jepsen believes that features she pioneered for the XO—such as the integration of the LCD screen and motherboard, allowing the CPU to shut itself down and save energy when little is happening onscreen—would benefit users everywhere, not just in environments where cost is critical or electricity is scarce.
Jepsen says she’s raising financing for the venture now, and that the first products based on the XO’s “holistic” design philosophy could hit stores as early as the end of this year. We caught up with her on last Friday, when she’d just returned from a whirlwind series of meetings with manufacturers at the International Consumer Electronics Show in Las Vegas. She commented on the prospects for a sub-$100 laptop in the near future, as well as the renewed acrimony between her two former employers—OLPC and Intel—over the giant chipmaker’s own attempt to market a low-cost laptop, the Classmate.
Xconomy: You’re starting Pixel Qi to pursue a new mobile-device design philosophy that you pioneered with the XO laptop. Can you talk about that for a minute?
Mary Lou Jepsen: I’m just back from CES, and I found it bewildering. Ninety-nine percent of the products are unnecessary. The iPhone and the iPod have redefined the high end of the consumer market, but nobody is doing that in the mass market.
If you look holistically at a device, whether it’s a cell phone, a laptop, or what have you, you can make products that are just as exciting [as the iPhone] in their own way—things that aren’t just for air-conditioned offices but that work indoors or outdoors, on or off the grid. But they have to be things that people are proud to own and proud to use. I think we accomplished that with the XO.
X: How will your designs be different?
MLJ: In order to work with economies of scale, we need to design a family of basic components that can be used by a variety of groups—not just children but adults, not just people in poor countries but people in rich countries. Everybody wants their batteries to last longer. Everybody wants to be able to use their cell phone or laptop or Blackberry outside in the sunlight and still be able to see the screen. It’s basic stuff.
Everybody thinks power is about the CPU, but the most expensive and power-hungry component in a laptop isn’t a CPU, it’s the screen. If you think of the screen as a separate component, not integrated with the CPU, then there is no way to lower the power consumption. But in the XO, the screen stays on, while most of the motherboard momentarily turns off while you are reading an e-mail or browsing a Web page or just pausing and reflecting on your writing. It starts up again with a key press. That is the basis of the power management architecture. That would never happen if you divorced the screen design from the laptop design.
I can’t help but think as a screen designer. As far as I’m concerned, the biggest feature of a CPU is how fast you can turn it on and off. People talk about low-power CPUs—well, you can get to zero milliwatts of power consumption if you turn it off.
X: Why did you decide to leave OLPC?
MLJ: I want to concentrate on the next-generation stuff, because that’s where my expertise lays. As much as I’d like to be an expert on education in the developing world, I’m not. I should be using my skills to the best of my ability to serve the cause. And the primary efforts of OLPC have to be on deployment and education and software right now. I’m happy to support those, but I just feel that I can help more by leveraging these economies of scale, not just with the XO but with many different products.
X: So forget about the $175 laptop, or the $100 laptop—you’ve talked about making a laptop for $75.
MLJ: Or less. I this that less is absolutely possible. I want to do it. I think that can be on the market by the end of 2009, with much lower power consumption than even the XO. And I think that lots of people would want such a thing. So the volume is there, but we have to bring the cost down. Not just by “putting a Dell on a diet,” as has been said, but by rethinking the design. For years, all of the component makers have marched to a drumbeat; every 18 months they have delivered the new components, and they snap together, and it’s worked incredibly well. But we’re at the end of that. It’s not about more megahertz anymore. The vast majority of people who have computers just need something to surf the Web and write letters and look at videos, and to do that you don’t need a gazallion megahertz.
There should be a dozen low-cost laptops on the market by the second half of 2008. But they’re all more expensive than the XO. Over time they should become less expensive, but I just don’t see anybody else working on that.
X: The news came out a couple of weeks ago about a pretty messy split between OLPC and Intel, with Intel leaving the OLPC board—they said—in protest over OLPC director Nicholas Negroponte’s alleged demand the company stop marketing its $299 Classmate laptop as an alternative to the XO. From inside OLPC, what was it like working with Intel?
MLJ: I was in Shanghai as recently as three weeks ago, working with Intel on their machine. And Nicholas has stood on countless stages advocating to people that they buy Classmates and XOs. On the technology side, we got off to a rocky start, but we were making progress. It’s hard coming off a “60 Minutes” expose, but once you get technology people working on a technical problem to find the best solution, you’re good. The issue was on the sales side, and the disparagement we were still getting from Intel. Peru was one really extreme example. Oscar Becerra Tresierra, the vice minister of education, said to me that Intel was coming to him almost every day and saying, “Look, we are on the OLPC’s board, and we know that the laptop will never work.”
X: Weren’t you working with Intel to put a low-power Intel chip into the XO?
MLJ: Yes. It’s very simple; Intel people want to sell Intel chips. I felt that the best way to work with them was to get some Intel silicon into the XO, which we were doing. The laptop we were making with them would have been more expensive and more power-hungry, and it would have taken at least six months longer to produce. But that was not the real reason [for the split]. The cost difference would have been less than 20 percent, and the power difference would have been maybe a watt or two, and we might have been able to pull a rabbit out of the hat and ship for mass production by the summer….[The problem was] on the sales and marketing side. I know how Intel is. I worked there. They are a very aggressive company and a very successful company. People describe them affectionately as the 800-pound gorilla of semiconductors, and that is what they are.
X: How fast do you hope to get some of your new designs into the marketplace?
MLJ: We are beginning to figure out what the next-generation laptop will look like. I’m hoping you’ll be able to see something in stores by the end of the year.
X: That’s fast!
MLJ: It’s a really fast design cycle, but with the XO we went from spec to mass-production-worthy prototype in less than six months.
Customizing the screen and integrating it with the motherboard was key to the XO. Nobody has thought of the display like a chip, but the manufacturing process for LCDs is that mature. The manufacturer had never taken an external design for an LCD before. I was the first one. But I didn’t screw it up, and I’ve got other ideas I’m ready to implement. The screen was just the first step.
X: Pixel Qi has gotten quite a bit of press already. Were you expecting that, and how does it affect your plans?
MLJ: I’m thrilled by the response. I didn’t really think it would capture this amount of interest. What it indicates to me is that it’s touched a nerve, and there is some pent-up demand. That’s good—I think we can deliver to that demand.
I’m raising financing and working on the design of the first actual products. I’m exploring many different paths, but my inclination is that we will sell products.
Everybody at CES was interested. I’ve been talking to some big-name companies that you’ve heard of. I am bringing together the investors and figuring out what the first products will be, and I’ll announce that when we’re ready.