In 2005, Mary Lou Jepsen joined Nicholas Negroponte, founder of MIT’s Media Lab, to build a $100 laptop for each of the poorest children of the world. As CTO of the One Laptop per Child (OLPC) project, Jepsen discovered that a laptop could be many times cheaper and more power efficient if the display was made differently. Thus, she designed a liquid-crystal display (LCD) that consumes only a fraction of the power of normal displays. And to ensure that it could be manufactured cheaply, she made certain that it could be built using existing LCD manufacturing technology.
Earlier this year, Jepsen cofounded Pixel Qi (pronounced “Pixel Chee”), a startup based in San Francisco that will make displays using her 48 patents on display technology. Next year, she says, displays from Pixel Qi will be on the market. Jepsen recently spoke at the Grace Hopper Celebration of Women in Computing conference in Keystone, CO, where Technology Review caught up with her.
Technology Review: Why did you leave the One Laptop per Child project?
Mary Lou Jepsen: My job was done. It was my job to figure out how to develop the laptop and convince the manufacturers to work with us. I decided that, paradoxically, the best way to help OLPC was to spin out a for-profit company and help lower the cost of the components that go into the OLPC. It’s economy of scale. If you make more of something, then you can make it cheaper.
TR: What technology lessons did you learn from OLPC?
MLJ: [That] it’s a lot faster and easier to use the large manufacturing faculties of the world as your lab, rather than a small little lab where you make handcrafted things, but you have to create the relationships and the structure to do that. One of the technology lessons was to work inside the cost envelope of the developing world to lower costs overall. What’s even more important and useful is dramatically lowering power consumption. Everyone wants batteries that can last 10 times longer.
TR: Why haven’t we seen much innovation in display technology over the years?
MLJ: A lot of people get really seduced by demos of the next display technology. I myself fell under that spell for about 20 years. I worked on heads-up displays, virtual-reality technology, and holographic displays–all sorts of really cool technology. It’s an emotional response to the display, and people want to have them. The truth is that over the last 50 years, only one display technology has gotten into mass production, and that’s LCD. There are two others, in smaller productions–plasma and DLP [digital light processing]–but they’re not in high volume.
TR: Your goal at Pixel Qi is to innovate within the LCD manufacturing process. How does this give us better displays?
MLJ: What became obvious to me after spending time at Intel [Jepsen was CTO of Intel’s now defunct display division between 2003 and 2004] was that the silicon people did things differently from the display people. Engineers who work with silicon send their design to a fab and have a chip back in months. But engineers making displays can’t just ship a new design off to a fab. So I thought, why don’t we just use the manufacturing infrastructure to get the high yields like the silicon people do? That’s what I did at OLPC. I went on to design a mass-producible product in six months, skipping the 20 years, millions of dollars, and missed window of opportunity that usually occurs in the display industry.
We’ve got deals in place with 40 percent of the LCD manufacturing industry. They said no to us initially. But we proved ourselves and our designs, and now they’re willing to make a bigger effort with us and our customers.
TR: How are Pixel Qi displays different from typical displays?
MLJ: We’ve got new screens based on the same ideas as the OLPC screens. Importantly, there are no manufacturing changes, and no material changes [compared with traditional LCD displays]; we’re using the same manufacturing process and following the same design rules. But what you can do that’s interesting when you change the design is produce sunlight-readable screens and super-color saturation. You can get really great reflective screens that rival e-paper at really amazing price points and with fantastic ultra-low-power capabilities. These displays have 1 percent of the power consumption of a regular screen by using a reflector behind the LCD grid to reflect ambient light and allowing the backlight to be turned down in bright environments. Plus, you can implement a power-management system used in OLPC [that refreshes the screen only when it changes] that saves you even more. We’ve done this by reinventing the display based on understanding the details of factories and how they work. All of these things work, and we are shipping them next year.
TR: Which products will they be in?
MLJ: We can’t announce our customers or products yet, but you’ll see these displays in low-power laptops.
TR: How do you see display technology developing over the next two years?
MLJ: I see an improvement in the readability of screens. The number-one reason why people print a page is resolution, and the number-two reason is that they don’t want to stare into a flashlight. Ultimately, in a year or two, we’d like to have a lawyer’s monitor or an editor’s monitor–some readable screen that’s just for reading. When I started meeting kids in the developing world and seeing that their schools were outside, we saw the opportunity to make screens more readable in sunlight. It’s ironic that the poorest kids in the world are getting the best screen technology through OLPC, but soon the rich people in the rest of the world are going to have access to it too.
TR: And beyond that?
MLJ: We have a road map that goes out pretty far. I mean, we can make improvements in all sorts of screens. Take the iPhone [touch] display. It’s actually two screens. One’s a touch screen and one’s a display screen. Why don’t we use the layers in the screen itself to do touch? It solves problems of alignment and lowers cost. We are following trends and watching how they evolve, and the great thing is that we can design a screen and have something that’s reliable that we can ship in about a year. We can actually do it in that kind of time frame.