Steve Perlman, the man behind tech incubator Rearden, is a serial entrepreneur with more than 30 years’ experience in the tech industry. He built his first computer from a kit during high school in 1976. He developed graphics at Atari and worked as a principal scientist at Apple, leading the development of a variety of multimedia technologies, including QuickTime. His string of start-ups includes General Magic, Catapult Entertainment, WebTV (sold to Microsoft for more than $500 million) and Moxi Digital. Since 2000, he’s been investing in new technologies at Rearden. The company, named after the company Rearden Steel in Ayn Rand’s “Atlas Shrugged” novel, is working on a wide range of R&D in a variety of fields and spins out start-ups as the R&D approaches commercializaton. Perlman believes that swinging for the fences, solving deep problems through years of research, and learning from mistakes is the only way to keep Silicon Valley moving forward. I was happy to have a chance to sit down with him at Web 2.0 and find out more about his business.
VB: How exactly does Rearden work?
SP: Rearden is an incubator. We’ve been around since 2000. What we do is take a look far forward into the future. We see a need for a technology or a production technique. We set out how to solve a problem. At that early stage, we won’t know how to solve it and won’t know how big the market will be when it’s solved. So, unfortunately, at that stage, you can’t bring in outside financing. We fund the companies ourselves. Most of the things we’ve tried have worked. A few don’t. But that’s fine. We learn from them. It very often takes years before we can show the technologies work. At that point, we build a large team and bring in outside funding. Then we move to product release. Some businesses stay within the Rearden fold and operate as cash businesses. We also cross-pollinate across all of these different companies. It’s not just ideas but people transferring from one company to another. A lot of the R&D that is gained from one company feeds into another.
VB: Can you give me an example of that cross-pollination?
SP: Ice Blink Studios did the art work for four feature films. That could be low tech. But those films were high-tech. They were “War of the Worlds,” “Monster House,” “Polar Express,” and “Beowulf.” We learned how to develop techniques for very large-budget films. We learned about the film business, like how they are financed, developed, distributed and sold. All that feeds into basic technologies we developed, like Contour Reality Capture, which is used to capture faces with our Mova start-up.
There were a lot of things we tried that worked for faces but were not optimized for the way people work on a production set. We had to learn about how almost all of the cost for traditional motion capture of faces and bodies for animations in films came from cleaning up the data that you captured. We had to make sure that whatever we did was as automated as possible in cleaning up the data. That seems obvious now. We came up with something that made it easier for the actors. We applied phosphorscent make-up to their faces. The cameras could capture every dot on the actors’ faces. It was comfortable enough. Now we don’t have to clean up a lot of inaccurate data. Then we talked to the video game folks, who are right behind the film folks in special effects. We bring high-quality stuff from films that is applied to games. Right behind the game people are web people doing Flash animations on web sites. These Flash animations are moving forward and you used to be able to only do them on a video game machine seven or eight years ago. You see a trailing effect. Rearden Studios does audio and visual editing. Mova does motion capture with Contour.
WOA.TV (Women of Action TV) is a community service site with videos of athletic women like Jackie Joyner-Kersee or Florence Griffith Joyner. But it is also a technology site. It was one of the very first sites with high-definition video being distributed on the Internet. As people used that site, we saw how well it ran on different machines. We looked at the algorithms. WOA TV was a complete test bed for us and a cool site for something that wasn’t covered enough, like women in sports.
VB: How do you bankroll all of this?
SP: I was at Apple early on. I did reasonably well there. I was also at General Magic. Did reasonably well there. Same for Catapult Entertainment. WebTV did reasonably well. I’m lucky in the sense that every start-up I’ve done produced liquidity. Given how many I’ve done and how many start-ups reach liquidity, it’s an unusual record. The ventures we’ve done within Rearden continue to feed back in and bankroll us.
VB: And how do you manage the funding for your startups?
SP: The companies are funded internally in early stages. As the companies evolve, the outside companies fund the [internal] companies, pick up the costs, and then they can move forward. They bring in capital, they operate, they are sold or do an IPO. We haven’t had an IPO yet. The Ice Blink Studios team became part of ImageMovers Digital at the beginning of 2007. ImageMovers Digital built their office around our office. Moxi Digital merged with Digeo. Rearden Studios and Mova are still part of the family. Other companies are unannounced.
VB: So, what brings you to Web 2.0?
SP: We’re interesting in telling the world about the technology we’re developing and how it relates to Web 2.0. A lot of what we do supports web-based stuff. More people are coming to us and saying they like our face capture. They say, ‘That looks great in a game, but I’d like to have it so we can use it to greet people at a web site.’ As computers are getting enough compute power in the processor or they have built-in graphics chips, our facial animation is becoming more relevant.
VB: Are there some trends that are guiding your future start-ups?
SP: We have a new web site. Go to Rearden.com. You see the areas of progress. Rearden Labs is the over-arching name for all of the future technology stuff we’re doing. We’re looking at anything relating to the web, with technologies at all levels from the server to the distribution to the client. We also look at content. We’re doing unbelievably cool wireless technology. That’s something we’ve been working on for a while. We’re doing cool stuff with imaging, including cameras, lenses, and image processing. One thing that fell out of that we’ve already released. That was Contour. Another area is alternative energy. We have developed some amazingly cool alternative energy technology. We have prototypes working there. Another thing new is biotech. I’m finally making my parents happy. They wanted me to be a doctor. My Dad used to say I should have studied. I’m finally making good.
VB: Why aren’t there many people like you doing R&D?
SP: Silicon Valley forgot how to do R&D. I firmly believe that now. I looked in every corner. There were three corners I thought it could be. One corner is in a large company like Apple or Microsoft. I was very senior in both of those places. Apple allowed me to do R&D when Steve Jobs had just left and our backs were against the wall. We did a lot of work in color Macs, sound technology, 3-D, animation, and QuickTime. A lot of stuff is still being used. By 1989 or 1990, we couldn’t get things through the system anymore. The Macs were selling well. People asked us to do smaller improvements. At Microsoft, I saw similar problems. They were focused on being competitive in different areas where they were being threatened. You’d figure start-ups would focus on something fundamental, like Intel did. That sort of product takes a lot of money and years before it can come to market. That kind of venture money isn’t around anymore. In fairness, the technology has matured and there are less risky places to put money. The last place to look was universities. They do it. But funding is scarce. You spend a lot of time looking for funding. The R&D isn’t done in the context of marketed products. They do basic technologies. I couldn’t find anyplace to do R&D. So I did Rearden. It wasn’t clear if this thing would take off back in 1999. It has. We’ve been able to produce companies that are producing revenue. We’ve had investments from VCs, financial investors and strategic investors.
VB: Are there any problems that are driving you crazy because they haven’t been solved yet?
SP: Let me get a list. (laughs). It bums me out that connectivity is not ubiquitous. There are dead zones. Wireless has bandwidth limitations. I’d love to create something with ubiquitous connectivity. I’d like to see simplified computing. You see a lot of that in software as a service. People are doing more and more stuff on the web and they’re installing fewer applications. Tom Paquin, who led the Netscape Navigator 1.0 team is our vice president of engineering. He’s very web-centric in the way he thinks. We totally connect. We would really like computer-generated characters, buildings and environments to be totally ubiquitous. We’re at a point now where 2-D animation and video are ubiquitous. Flash has that covered. If you want to do 3-D now, it’s complex, expensive and a tricky area. Some computers can handle it and some can’t. We’d like to address problems like that.
VB: How many researchers are there within Rearden?
SP: Rearden ebbs and flows. We have somewhere between 50 to 150 people working on different things. Between projects, we have small numbers. We go low when a company gets acquired. People move over. We focus more in one area and less in another in different times. WOA TV was once a very focused project. I can’t tell you how many iterations we went through to get that right for the web in 2003. Trying to get HD streaming through the web five years ago was unheard of.
VB: Do you feel any kinship to Nathan Myhrvold’s Intellectual Ventures? One thing he does is hire a smart person before they’ve invented something.
SP: I appreciate that. There’s a germ of a similar spirit there. But he’s doing a lot of stuff with patents that I don
’t understand. The patents Rearden files are entirely for our own use. When we build something, we file patents as we build it. We don’t buy patents like they do.
VB: You actually made some headlines by coming out and standing up for small inventors on the legislation being proposed to overhaul the patent system.
SP: The issue with that bill is that it ended up being very politically driven. No inventors testified. They didn’t take testimony from a single Silicon Valley start-up. We came in and said what is going on here? We said the bill would harm our ability to do start-ups. They didn’t listen. Fortunately, the National Venture Capital Association spoke up. They say they want patent reform. Where are the inventors? Everyone realized the emperor had no clothes. There never was a final draft of the bill. That seems to be where it’s ended. But we’ll see.
VB: Safe for now?
SP: Yeah. People don’t realize that patents are the currency of innovation. It’s not to say they’re perfect. But they’re all we’ve got. A lot of what we do is simplify something to where it can be easily replicated. If it can be easily replicated by a large company jumping into our space or a country with a low cost of manufacturing, we can’t compete. Rearden from the get-go was a high-risk proposition. If you take away the patents, it becomes an infinite risk proposition. Someone else can just replicate what we’re doing.
VB: What’s your conclusion about taking big chances — swinging for the fences?
SP: It’s the only thing worth doing. Everything else is booorrrrinng. The second thing is it’s the only thing that moves the ball forward. I love incremental work. It’s terrific we’re doing it. Rearden Studios invents very little with audio and video editing. That’s fun. But we aren’t going to have any world-changing things if we don’t allow people to take chances. That got me really depressed as I went to the large companies, start-up companies, and universities. Where do you go next? That’s what Rearden’s for.
VB: How much of the idea development are you personally getting your hands on?
SP: All of it. The one place I didn’t make a personal contribution to was Ice Blink Studios. That was Doug Chiang all the way. The guy is a genius. What I was able to help Doug out with was the business structure. We did incubation. We helped with the web infrastructure, financing and other stuff they couldn’t do.
VB: So you leave yourself enough time to be directly involved in every project?
SP: What I’ve gotten good at over the years is sticking my nose in when I’m needed and keeping my nose out when it’s in the way. Sometimes I serve as a counterpoint to them. It forces them to think the right way. I’m like a muse. Sometimes I take someone with a deep knowledge in one area and give them a broader context. Either in a different area of technology or a creative context or a business context. A lot of guys agonize over cost-reducing something in the short term when they know they can do it in the long term. I tell them not to sweat the short term. Get it working. As we march to the market, we can cost reduce it afterward. This may be throwaway work as you find things later that can bring the costs down. Other times, maybe the end result is a consumer product. And the whole point of the product is to make something less expensive for consumers. I can give them a business context.
VB: Is there anyone you admire who’s able to get things done quickly?
SP: I like Android, which I had an investment in. It’s a platform on top of Linux that goes beyond Linux. They’re applying it to phones initially but it can go beyond that. I’m very curious to see where they take it. I’ve seen prototypes. As for other cool stuff? Sometimes I see cool things but don’t think it’s a good business. I think Adobe’s Flash is going in a good direction. I think it’s exciting. I’m curious what will be done with Adobe Air. I think we’re past the age of applications now. We’re moving to this area of cloud computing or applet computing and how those different things merge together. I think people offering particular services or middleware or packages you can build your services on top of, that’s all cool. A lot of our R&D is devoted to that area.
I was never successful at getting people to develop a real-time operating system. The operating system often gets in the way of things like games. The operating system does things like manage how resources are distributed equally, whereas a game wants all the resources. If you’re trying to play Crysis, you just want frame rate. You want the graphics to look good. You don’t want the game to pause. The games just shut the OS down. With Unreal Engine and Crytek, you see a new generation of code for which time is everything. The Flash player is like that too. It thinks from the get go about time and the real-time delivery of things. I want to get past waiting for the OS to catch up.
VB's research team is studying web-personalization... Chime in here, and we’ll share the results.