Philippe Kahn is back – with a company that he just can’t stop saying is going to change the world.
Called LightSurf, that company is an infrastructure play – albeit for a world that doesn’t quite exist yet. At its surface is a camera the size of a keychain charm that snaps onto a cell phone and transmits a photo over a cellular network directly to the Web, an email file, or traditional paper photofinishing at the click of a button. So far, LightSurf works only with Motorola’s Accompli line of cell phones, to be introduced in December (the combo will sell for around $500), and only on the Pacific Bell network. But that’s OK by Kahn, because other customers will come, and while the camera’s cool, it’s the supporting architecture that really matters. “If you go back to what I refer to as the go-go years of the Internet, you could throw money at anything and it would work,” he told a crowd of 300 CEOs at last May’s Vortex conference in Dana Point, California. “We went to business-to-consumer models, and then to business-to-business models. These are fads. But one thing is absolutely certain – the Internet will grow, and infrastructure will grow with it.”
LightSurf’s infrastructure includes everything involved in the ephoto process, as Kahn calls it – from the camera’s embedded software to the server technology, from a new image-acceleration scheme designed to work over wireless networks to a massive storage and filtering system that will house billions of photographs. Where others see a tiny camera piggybacking on a phone, Kahn and his team see an end-to-end solution powered by LightSurf – “a Cisco-sized opportunity” that will allow parents to send instant photos to Grandma, help ordinary citizens deter crime, and, hell, further the cause of democracy while we’re at it. To Kahn, nothing is outside the realm of imagination, because, in a world where anyone can instantly publish pictures on the Web, everyone becomes a photojournalist. And in Kahn’s view, a camera could be our most powerful weapon.
The company’s first two clients, Motorola and Kodak, pay LightSurf to develop, deploy, and manage new ephoto products. Kahn says the next step is a Qualcomm-type model, where LightSurf gets a cut of every device sold. But he’s not stopping there. LightSurf’s network infrastructure is designed to allow the company to collect a fee every time anyone snaps a photo. “Our ambition is to put a visual communicator in every shirt pocket, on every belt, in every handbag,” Kahn says. “LightSurf is an operating environment for instant wireless digital photography.”
No matter how grandiose Kahn’s vision may seem, it has some grounding. Consumers do want to share photos online; in a 1999 study prepared for AOL by Roper Starch Worldwide, 86 percent of those surveyed said as much. Photo sharing was the most requested service – surpassing the desire to take classes, pay bills, or even download music online, and sparking AOL’s “You’ve Got Pictures” initiative.
But here’s the thing: Technically speaking, sending a photo from your desktop to Uncle Dave’s is a walk in the park compared with transmitting it through a cellular network. For starters, cellular service drops out all the time. And even if the file goes through, the top cellular-data transmission rate in this country is 14.4 Kbps. Transmitting a 100K VGA-resolution photograph would take a few minutes. That’s too long, and besides, any appreciable volume of photos would hose a cellular network.
Of course, Kahn has never had much of a problem with technological hurdles. He’s always lived close to the machine. In the 1970s, after receiving the equivalent of a PhD in mathematics, in France, Kahn worked on Pascal under Niklaus Wirth in Zurich and was a programmer for the Intel-based Micral, one of the first personal computers. He arrived in Santa Cruz in 1982 and created Turbo Pascal, Borland’s first big hit. “Philippe Kahn is a prodigious software developer,” says longtime industry observer Dave Winer. “He really blew some holes in the notions of what software could do. One of his first products, Sidekick for DOS, would let you toggle back and forth between, say, your spreadsheet and your address book. That functionality showed up on the Mac only years later, and then in Windows.”
This time around, Kahn is breaking ground with a class of mathematical functions known as wavelets. It’s a big move away from the status quo in digital imaging. Digital photography has traditionally relied on a standard known as JPEG, whose mathematical underpinnings date back to the 18th century. A typical digital camera compresses an image by turning the tiny changes of color and light in a photo into coefficients that an image software package can interpret using cosine formulas. The problem is that cosines can describe the frequency and degree of changes but they have no way to describe location, and so can’t focus on areas of abrupt changes.
In the early 1980s, mathematicians constructed wavelets as a way to better represent rapidly changing data – like those caused by earthquakes, nuclear events, or quick shifts in color and light. Wavelets reflect the frequency, degree, and location of big changes in a data set, and can describe images on various scales: Some wavelets represent the whole image in a crude sweep; others pull out details. In 1987, a Belgian mathematician named Ingred Daubechies helped take wavelets a step further when she began writing the algorithms that would encode digital images using wavelets. A decade-plus later, Kahn will be the first to commercialize this work.
Kahn came up with the idea for LightSurf three years ago, when he tried to send an ephoto of his newborn daughter to her grandmother. While many new dads would be looking around for a shave and a stiff drink, Kahn was crawling the maternity ward for an RJ11 jack. With no landline in sight, he built a crude wireless photo application in Borland C++ and sent the pictures over a 2,400-bps connection from his StarTAC. Later, figuring there had to be better way, he brought together a team and set out to design a way to send a slender image file for online use and a larger one for printing hard copy. It worked just as he imagined. Meanwhile, another group of scientists was working with wavelets to develop a new digital imaging standard, called JPEG 2000, set for commercial release early next year. When Kahn became aware of the development, he worked to make LightSurf’s format compliant with JPEG 2000.
Following the path of an image in JPEG 2000 shows how much bandwidth wavelets can save. As soon as the camera’s 640 x 480-pixel sensor captures the image, a Samsung ARM Core chip begins crunching a photo into bits. LightSurf’s software, called PhotonOne, then hands the data – representing an image file about one-third the size of a typical JPEG – to the phone as TCP/IP packets. Kahn calls them half-baked mathematical expressions waiting to get fully cooked at the server.
Kahn is most proud of what happens once the file arrives at his network. His programmers have tweaked the Linux source code to handle file acceleration and implement the second major component of the business – the LightSurf Universal Imaging Switch (LUIS). This switch, which sits on the network, reads the resolution, file format, and intended destination of an image, then routes it to the appropriate server. “LUIS does for images what Cisco routers do for packets,” says Kahn.
“Our ambition is to put a visual communicator in every shirt pocket, on every belt, in every handbag.”
Customers never have to contend directly with JPEG 2000 – they need only know where they want their photos to go. If print is the destination, the server sends the highest resolution. For the Web, a medium resolution makes transmission faster. A Palm VII grayscale takes the lightest load. Kahn calls it all a “massively scalable infrastructure” for storing, managing, securing, and serving up millions of images over time. In theory, the entire process – for a photo that will look great in any medium – should take no more than 30 seconds. In practice, as of late July, LightSurf had some work to do. The first samples looked great online, but were too pixilated in print. Nevertheless, they arrived in time.
LightSurf is passing its first market test seamlessly. Since June 26, Kodak, powered by LightSurf, has been enabling consumers who drop off film at any one of 4,100 CVS pharmacies to have their images digitized and stored online. When the film is developed at Kodak’s Qualex labs, Kodak scans it at a 1,000 x 1,000-pixel resolution and sends the images to LightSurf, which delivers them to a password-protected Web page. For the customer, it means no more keeping track of negatives; for Kodak, it’s all about new print orders. As the enabler of it all, LightSurf gets a licensing fee plus fees based on usage.
It’s only a piece of the full package he wants to provide, but Kahn estimates that LightSurf handled more than a million photos in the two weeks following the project’s debut. And Kodak couldn’t be happier. “Philippe always has some innovation going,” says Dennis Hamann, Kodak’s general manager of consumer imaging, who collaborated with Kahn while in the PDA business at Hewlett-Packard. “He’ll bring the technology to companies like HP or Kodak that can’t be as innovative but have the brand recognition. That’s why I chose to work with him.”
These days, Kahn has three homes, five sailboats, two jets, and five surfboards spread between Santa Cruz and Honolulu. His wife and LightSurf cofounder, Sonia Lee, is a graphic designer by training. They work together on their philanthropic Web site, the Lee-Kahn Foundation (www.lee-kahn.org), as well as on LightSurf and Starfish, where Lee’s duties encompass everything from serving as a sounding board for her husband to functioning as VP of human resources. LightSurf’s sole investors, the two are living well off the $253 million sale of Starfish to Motorola in 1998 (Kahn is still CEO).
But while Kahn’s vision has afforded him a nice lifestyle, there’s one thing he’s lacking: a demonstrated ability to build the kind of business he’s talking about. While Starfish offered some amazing sync technology from the start – linking the Palm to Microsoft Outlook – these days it’s a niche product used mostly to integrate PDAs with Web-based organizers. Kahn also had grand plans for Rex, a PDA that slid into a laptop’s PCMCIA slot and could fit in a shirt pocket. He billed it as wearable tech; but Rex went nowhere – a fact Kahn blames on his decision to license the device to Franklin Electronic Publishers, a New Jersey company he claims fumbled the marketing. Just as likely, Rex simply didn’t have the functionality or appeal of the PalmPilot, which came out a year earlier.
The closest Kahn has come to a LightSurf-sized idea was with Borland. From the outside, Kahn appeared extremely involved in the company’s success – and failure. “He was always very hands-on, very robust, and very rigorous,” says former Intel chair Andy Grove. To insiders, though, Kahn seemed chaotic: He was a programming whiz who’d spend hours suffering over ad copy. And as a manager, he’d motivate the hell out of his staff to build world-class technology, say he wanted to be involved in every decision, and then disappear for days, or weeks, at a time.
To this day, Kahn still wants to do everything himself. “A lot of people have been knocking on our doors, and we have been pushing back. I think succeeding is learning how to say no,” he says. “We are a technology provider, an infrastructure provider, building all the pieces.”
But Kahn has grown. Those close to him say he’s learned a lot about management, and for whatever his faults may be, he knows how to get things done. “From time to time, people have sold Philippe short and have come to regret it,” says Gary Reback, CEO of voice-over-IP startup Voxeo and Kahn’s former lawyer in Borland’s successful defense against a copyright suit by Lotus. “The thing he really does well is create new companies with great ideas and incredible people.”
Kahn has also learned the importance of strong marketing partners. In fact, in its desperate race to catch Nokia, Motorola may be the perfect ally. While Kahn designed the camera, redesigned the V-series phone to work with it, established the server farms, designed the software, and even set up photo printing facilities in the LightSurf offices with a giant $250,000 Konica photofinishing lab, he did it technically as a Motorola employee – and completely in lockstep with Motorola’s vision of the wireless future. “We needed an end-to-end solution, so we can show people what we’re trying to achieve and what to expect going forward,” says Jonathan Ruff, Motorola’s director of business development. “We’re looking at this as something that will help handset sales and demonstrate leadership in the market. Digital camera adoption is really a sweet spot right now.”
If the Accompli takes off, no doubt Nokia will also want some of what Kahn has to offer. And then there are vertical opportunities on the back end, such as in the medical arena, says Robin Nijor, Kahn’s marketing adviser. “If there’s a doctor in a small hospital at 2 am with no radiologist, it would be awesome to know he could send a CATscan to London, where it’s prime time.” Or in law enforcement: “When a suspect’s picture comes into the station, it could be sent out to squad cars with Palm VIIs.”
If LightSurf doesn’t pan out the way Kahn envisions, it won’t be for lack of commitment. Andy Grove says Kahn’s most impressive trait is his belief in himself. “Philippe may have gotten ahead of himself at Borland and built a big building. But he thought he was going to win. I admired that about him. They broke the mold when it came to Philippe.”
Whatever happens, Kahn is definitely back, and with a huge idea. How can he pull it all off? It’s like asking how anyone can run three companies and still have time to be an accomplished musician, much less surf. For most of us,it’s impossible. For Kahn, it’s just another day at the beach.
How will the world be different if LightSurf succeeds?
The Internet becomes less a consumer thing, where we are now, and much more of a place where people communicate. In other words, the original vision of the Net, before commercialization. What will happen is that suddenly we’re going to have a lot of content created on the Internet – visual content. With this kind of device, you’re going to see the best and the worst of things.
The first camera-phone combo, from Motorola, will retail for about $500. How many will sell at that price?
That’s where it’s starting. What’s unknown is what kind of subsidies and programs are going to hit once people get excited. If Verizon or Pac Bell just goes gaga, who knows what the price will be with subsidies – maybe $199?
Why didn’t you hire an industrial design firm to design the camera?
It was key to hire guys on the inside who could make sure we stayed within the realm of what was doable. When we designed the external part of this camera, there was no way that all the parts, all the electronics, the sensors, the battery life, would work in such a small space. But the beauty of having internal guys who understand how these things evolve is that we knew it would take 18 months to build, and we had faith we’d be able to fit all these parts in by the time we shipped.
You created the back end, too.
A complex engineering solution. We had to make it all work so that when you click, you order your print. It looks like magic, but there’s a lot of work that needs to happen for a quality photograph to move through the air. And we see ourselves as a provider of that service to a lot of players, not just to Motorola and Kodak.
Who’s the competition?
My Sony Vaio C1 has a little camera, and if I tie the laptop to a Ricochet modem, you could argue it’s competition. But it doesn’t have the back end. That doesn’t mean we can’t build a back end for this; we can. But we don’t see any direct competition.
Why didn’t you go the VC route?
I think that if you need venture capitalists, great. If you don’t, you may be better off.
The venture capital community in the last few years has gotten very greedy in wanting very rapid returns. When the VC is looking at his watch, and his return is not coming, the vision starts to change. CEOs change, everything changes. And the reality is, build something innovative that solves a difficult problem and you will have something. Most of the long-term successful companies were built on those terms. Certainly Apple. Sun is another example. And in some ways, Motorola has survived because they are willing to make mistakes. I loved Iridium for one reason: It was a massive risk. It’s easy for people to say they were idiots down the line. But you know what? They took the business risk, and fine, they got their clock cleaned, but they had the vision to take a leadership position and launch those satellites. Most companies don’t have the courage.
So the Valley’s changed for the worse?
In many ways, Silicon Valley is not a technology valley, it’s a marketing valley. A lot of the startups you’ve seen in the last few years have been great marketing twists or business twists, people who were there to make money first, innovate second. My personal approach is very different. It’s, let’s innovate, and as a consequence of that, make money.
In that spirit, you used Linux source code as the base of the LightSurf universal imaging switch.
From a financial perspective, Linux is a great thing, because you can’t beat the cost/performance. And Linux is at a stage where you can actually use it to build a massively scalable operation like ours. Three or four years ago, if we had gone to Motorola and said, “We’re going to build this back end and we’re going to use Linux,” they’d say, “Go buy Solaris boxes.” Today, Linux is accepted, and the beauty of it is that we can actually enhance whichever implementation we have without making a proprietary system.
Do you consider wireless the most exciting aspect of technology right now?
If you’re a young kid today and you’re told to work on a word processor – well, why would you want to do that? Right now, young kids want to build wireless devices and cool things. I don’t blame them. That’s what I want to do, too.
Who’s the Microsoft of wireless?
There is one company that a lot of people point to – Phone.com – and say they’re trying to control WAP and make it nonstandard. But I think they’re sincere in wanting to make the mobile world work together. There is no Microsoft here because there’s so much competition. I think the best thing that happened to Motorola is Nokia and vice versa. The best thing that happens to consumers is these guys competing for business, which lowers the costs of services and handsets.
To the point where you buy a new phone every six months just because it’s cool.
Exactly. There’s a feeling I get with a new toy or a new pair of shoes. I love getting new shoes. You always have one or two good days – wow. They feel good on your feet. Getting a new phone is the same kind of deal.
Borland is still around – sort of – but only as a shadow of what it once was. What happened?
The challenge we had at Borland was not about management; it was about the competitive landscape and Microsoft, with FoxPro, willing to give away what we were selling. I don’t want to defend things, but Borland systematically innovated. Unfortunately, our products were all competing with Microsoft products, and they were willing to give them away as long as they had to – exactly what happened at Netscape. Are we going to say that Jim Barksdale didn’t know what he was doing, that he was a bad manager? The problem is, he got into an untenable position, where everything he was trying to sell was given away by Microsoft, and at the end, he had to sell the company. That was tremendous vindication for me. Was I perfect? Absolutely not. But we weren’t total idiots either.
Didn’t you use the same strategy – underselling dBase with Paradox – back when you were competing with Ashton-Tate?
But we were pricing it to be competitive – not in a predatory way. We weren’t out to destroy the category. We weren’t saying, “Hey, we’re not really a player in the database category, but we’re going to give away Paradox until Ashton-Tate’s in trouble.” Microsoft was basically giving away what Borland was trying to sell. It’s hard to compete with a company like that.
Do you think Microsoft should be broken up?
I don’t have any direct opinion on that. I hate government. Although Microsoft was a direct competitor, I don’t like government to run things.
But do you think a Microsoft breakup would be good for users?
Well, one thing is true, if I look at the state of spreadsheets since Quattro Pro, it hasn’t changed; in fact it’s worse. There needs to be innovation there. Microsoft is an entity – it is in some ways an old-world company. But it has a lot of power, because everybody uses a word processor or a spreadsheet. If you set your sights on building the best possible text processing system, you would never build something like Word. But if you set out to have a monopoly on word processors worldwide, you’d probably build something obscure and complicated that people can’t reverse-engineer or clean-room-implement. When you have 90 percent share, you tend to go to 100 percent, not 80, because it’s harder and harder for people to compete. I think it’s a fundamental problem. Is the best solution to break up Microsoft? I don’t know. I’m lucky enough to have lived in a non-Microsoft world for the last six years. Microsoft is not a factor here.
It’s been a while since you’ve had a big hit. Is LightSurf going to be it?
I don’t think that way. With LightSurf, we said we better find somebody that we can work with. And the two leading contenders were Motorola and Nokia. Motorola has 100,000 people worldwide and a global operation. I think that gives Motorola some leadership. We’re looking for global leaders that have a great marketing channel and can push the brand. We’re in the “powered by LightSurf” business. We’re not in the business of creating LightSurf as a consumer brand.
What’s your home life like these days?
It’s absurd. The house is out of control. We have 11- and 16-year-old nephews, Sonia’s nephews, who basically live at our home. They kind of moved from Michigan and decided they liked it. They’re all into programming and stuff. And now they’ve multiplexed the hubs. Right now they’re on summer vacation; still they’re up at 6 am. I’ll go to the bathroom, and I hear laughing and stuff like that, at 6 am. They call me Uncle Taz. “Uncle Taz, you there?” And I say, “You guys are still up?” It’s absurd. It really is.
And you have more kids, too.
I also have a daughter who is 26 and a daughter who is 24. They are both biotech engineers, interestingly enough. My older daughter works at a startup, to pay for her PhD studies. I have pretty strict rules with my kids. They have to make their living. The trouble when you succeed in business or whatever is, you become known as a successful person and it’s hard for your kids. I know for my daughters it was tough. And so I made it even tougher by saying, “I’ll pay for your studies, but that’s all you get.” The toughest thing in the world is to raise kids. And it’s funny, because right now I have some in university, and Sophie is in diapers. I want my kids to grow up and see innovation, just like I grew up. My dad was an aerospace engineer who designed the wing on the Concorde. It’s amazing what they did without computers. I saw my dad working long hours, perfecting a design.
So what time do you get up in the morning?
It depends what time I went to bed. Monday I was up at 6. Other days I’ll be up at noon because I was working until 6 am building some software model. It depends on x iteration of this, or whatever I have to work on. Nights work pretty well, no one bothers you, and you just work through the night, and at the end, you got something. It’s the magic of the night.
What about free time? Do you still play music?
Every day. Music is kind of a religious experience for me in the way that I practice religiously stuff that I know I’ll get better at and then I’ll die. My dad taught me music, and I’m teaching my kids. Studying music is important because of the discipline it requires. If you don’t practice every day, there is no way that you’re going to master an instrument.
Why Santa Cruz?
When I got here in ’82 or ’83, when I came from France, I came on a tourist visa to find a job as an engineer. And I ended up working as a consultant, because I had no green card and nobody would hire me. And one weekend, one of the customers said, “Oh, we have a beach house in Santa Cruz. Why don’t you come and check it out?” I came here. I said, “Why work in San Jose?”
Now you’re legit, because you surf.
Surfing is a pure sport. It’s wonderful because it’s such a simple thing: the board, the ocean, and you. Surfing is definitely one of the things I wish I had discovered as a kid. Because it doesn’t cost anything. It’s pure skill. It’s an incredible workout. And we are so close to the ocean. But I am happy to have discovered it in my forties rather than never.
What kind of board do you have?
I have a long board made and designed by a local guy, Johnny Rice – a famous surfer and shaper. Johnny must be nearly 70 years old. And he lives by the ocean. He made me a huge 12-footer that’s fantastic, catches any wave.
[Kahn’s wife, Sonia, arrives.] How is it working with your wife?
Sonia’s a facilitator. Whenever there is something to resolve, she’s great at helping to find some resolution. We’re a team. Sonia and I are friends as much as anything, and partners.
Sonia, what’s your take?
Philippe always says, in regard to relationships, it’s not how you argue as a couple, it’s how you resolve it. And I think that’s the same when you work together. I think going through that Borland saga has made us stronger. You go through certain crises in your life, and you come out of them. And either they will hurt your relationship, or you end up coming out stronger for them.
Did the Borland saga make you a better manager, Philippe?
One thing that I remember when things were getting tough, Sonia said: “There’ll be better days.” That was perceptive, and helped get me through. But there were some tough days. That’s probably part of the management process, one of the things you learn. There are a lot of fair-weather managers. It’s easy to be a great manager in a company that’s growing. It’s much harder when things are tough.
They say sailing, which you’re doing a lot of these days, is a good metaphor for business. True?
Sailing is tactics, strategy, patience, and persistence. And lots of technology – understanding the wind, the storms, the weather maps. You have to be smart and have stamina. There are the parameters you control and others that you don’t. For me, that’s a great challenge. And it’s the same when you launch a business. You try to pay attention to what you’re creating and the people you hire, but there are things beyond your control – like what the competition’s doing. With a startup, there are no shortcuts. We’ve spent years developing the technology, and there are days you look at it and say, “Oh, geez. Are we sane?” But it’s a distance race. You can’t worry too much. You have to commit.