Incorporated in Aibo, the cute, silver-toned pet robot dog that went on sale June 1, was much that resonated with the history of its maker, Japan’s Sony Corporation.
For instance, the novel coupling of “entertainment” and “robot” recalled the pairing of “personal” and “stereo” in the Sony Walkman two decades earlier.
There was the “Memory Stick” lodged under Aibo’s tail, a new data storage medium about the size of a piece of chewing gum that could someday become as ubiquitous as that other Sony invention, the 3.5” floppy disk.
And Aibo’s rechargeable lithium-ion batteries were yet another Sony original, developed in this $50 billion giant’s seemingly never-ending quest for smaller, faster, friendlier consumer electronics products.
Those innovations fit neatly into Sony’s track record. But Aibo had at least two unfamiliar aspects. One was the way Sony chose to make the new home-entertainment robot available: only over the Internet, where all 3,000 units of the Japanese allocation were snapped up in 17 seconds. The other was Aperios, its operating system.
Wait a second… an operating system from Sony? What gives?
The digital revolution. Just before the final Christmas shopping season of the millennium, the world’s best-known electronic consumer goods maker has decided to re-invent itself for the Internet age. The transformation is being led by Nobuyuki Idei, the tough-minded president who took Sony’s reins in 1995 and took to the digital agenda in a big way. The new motto he gave the company, “Digital Dream Kids,” is also an excellent description of Idei’s unusual brain trust-the Sony Computer Science Laboratory (CSL).
Founded a decade ago, CSL is a very un-Japanese research shop set up in emulation of the mother of all computer science labs, Xerox’s Palo Alto Research Center (PARC). At CSL’s Tokyo offices, thirty-odd researchers now work on deep concepts of computer connectedness and far-out interfaces. What’s more, they’ve been tapped by Idei to help lead him and the rest of Sony’s 170,000 workers into the new world of “convergence”-where the PC and home audiovisual appliances merge, and Sony battles Microsoft.
In fact, that battle has already begun. The Aperios OS, invented at CSL, and Microsoft’s Windows CE are squaring off in a struggle for preeminence inside the TV set-top box, the latest portal for digital cable services and a possible linchpin for the high-speed “home network” that Idei’s company hopes will ultimately link each of its products not just to one another, but also to the Internet.
Not Sony’s Way
Since the company’s founding in 1946, Sony has largely been driven by physicists and materials specialists whose style of innovation was to take key enabling devices and turn them into new and useful consumer products. A classic example of Sony’s eye for hardware’s “killer app” is the transistor, which Sony turned into its first hit product, pocket-sized radios. And it was Sony (after spending 13 years and $200 million) that first transformed the charge-coupled device (CCD) into a marketplace success-in the form of its best-selling camcorders.
Although wildly successful, Sony’s pattern of innovation did not extend to computers. Sony management had never put much effort into computer R&D, believing the company’s core business would always be manufacturing groovy audio-visual equipment. The strategy posed a stark danger: that Sony would become irrelevant in the emerging digital world.
That danger was particularly clear to Toshi T. Doi, an expert in digital audio who had played a key role in the development of the compact disc player. Doi arrived at Sony’s computer business group in May 1984, just in time to witness the launch of Hit Bit, a home PC that he wryly describes as “the last eight-bit machine in the world.” Other failures followed, but undaunted by Sony’s computer allergy, Doi and a group of 11 engineers proposed developing a serious, low-priced engineering workstation called NEWS. “It was our last card,” Doi recalls. “If it had failed, that would have been all for Sony” in the computer area. Launched in January 1987, NEWS was a big hit in Japan, quickly taking top share in the local market (though it tanked in the United States, bested by superior products from the likes of Sun Microsystems).
Though NEWS was successful, Doi was dissatisfied. “When I opened the box,” he recalls, “it was nothing but assembly-we were buying CPUs from Motorola, licensing Unix from UC Berkeley. There wasn’t any Sony original technology included, and I thought, ‘This is not Sony’s way.’” Where, wondered Doi, was the innovative technology core that could drive new products and markets, just as the transistor and the CCD had?
Working with Unix had alerted Doi to the importance of the Internet, and he foresaw that in the 21st century, networks would connect billions of computers. In April 1987, Doi asked his management to let him establish a software lab to focus on network design. NEWS’ success made Sony management receptive to Doi’s proposal. He got funding and a new mandate: Think long range.
As he began putting his group together, Doi had PARC very much in mind. During the 1970s, PARC was a place where researchers enjoyed unique freedoms. It had proved an excellent research strategy. From PARC came the graphical user interface, the laser printer and the Ethernet. To get his project up and running, Doi asked his friend and fellow amateur jazz musician, former PARC luminary Alan Kay, for guidance. At the same time, Doi approached Mario Tokoro, a Keio University professor known for his strong technical background in operating systems and networking, as well as for his visionary streak.
Tokoro remembers that Doi’s timing was perfect. “At that time, as a professor, I was a little frustrated,” Tokoro says. “The PhD students I produced were not given a chance to fully exhibit their performance even in the research labs of top-level Japanese firms. Bureaucracy and strict seniority systems were so strong. Originality was not considered important. And I was thinking whether I should start something.” The pair saw eye to eye-networks were the future, and intellectual freedom could make Sony a player. Doi offered Tokoro the job of lab director.
Freedom Breeds Content
These days, a good place to meet Tokoro is in the massive self-service cafeteria at Sony’s Tokyo headquarters in Shinagawa, where he lunches most days with colleagues from CSL. Tucked inside the southern tip of the Yamanote, the loop line that defines central Tokyo, Shinagawa could these days justifiably be renamed Sony Town. The company owns at least a dozen high-rise buildings in the area, and its logo is everywhere. The CSL is located in an eight-story tile and glass edifice hidden on a side street, out of sight of Sony’s headquarters complex, but no more than a five-minute walk from president Idei’s office.
CSL’s director is short, pudgy and dressed in a standard-issue salaryman suit; a casual encounter yields few clues about what lies behind the conventional facade. A better insight can be found on his Keio University Web page, where one of his students mischievously posted a cartoon of him clutching a Nintendo-like magic mushroom. The implication is obvious: shy, mild-mannered Professor Tokoro is actually… Super Mario!
In Japan, where universities tend to be ivory towers and academics are suspicious of industry, it is extremely rare for professors to join companies. Indeed, Tokoro made a lengthy transition. For the first 10 years of the CSL’s existence, he kept his position at Keio and worked for Sony part-time. In March 1997, he joined Sony as a full-time employee. Two years later, Tokoro was promoted to senior vice president making him, at age 52, Sony’s most senior corporate research executive. “That’s a very strange place for an academic,” comments Dave Farber, a University of Pennsylvania computer science professor who has known Tokoro for many years. “There’s a president [Nobuyuki Idei] and two people under him, and Mario’s one of them. And if you look at Sony, the new president has essentially said, ‘The IT area’s our future.’ So that puts Mario in charge of one of Sony’s most important commodities.”
Inside the CSL, the atmosphere is casual and relaxed. Researchers wear polo shirts, jeans and sandals. A sticker on one door reads “MIT Nerd Pride.” The lab is decorated in muted colors, with central meeting areas furnished with soft black sofas. Spacious individual offices take advantage of the light from surrounding windows. There is no obvious center of power, no signs indicating titles or seniority. It is, in short, completely unlike most Japanese industrial labs.
“Much of CSL’s uniqueness comes simply from the fact that it’s essentially a very Western lab,” says Rodger Lea, a 36-year-old Englishman who worked there from 1994 to 1997, and now runs a Sony software development lab in Silicon Valley. “Japan is such an immensely conservative, hierarchical bureaucracy, whereas CSL is a meritocracy, people are successful based on their ability to perform.” Unlike the rest of Sony, CSL employs researchers on a one-year contract basis, with annual performance reviews. Salaries are high, but they are based on success, not on seniority.
Not surprisingly, there’s no shortage of applicants hoping to join Sony’s digital dream team. CSL has attracted some of Japan’s most promising computer whizzes-including some from other companies. Until recently, job-hopping was a rarity in Japan, but CSL’s offices are filled with renegades from the likes of Canon and NEC. All must meet Tokoro’s definition of a good hire: “Originality and vision; the strong will to deny orthodoxy and to challenge the future.”
To promote the creativity that produces good software, Tokoro adopts a Western, hands-off style. “His policy is that the best management is not to try and manage,” explains Hiroaki Kitano, one of CSL’s most productive researchers (see sidebar: “Sony Supernova”). “All he does is hire the best researchers and let them do what they want. In five years, he has never told me what I’m supposed to do. He always says, ‘Do whatever you want, and the only thing is to be the best and have results with the highest impact.’”
Network of the Future
Despite CSL’s freedoms, Tokoro is well aware of the pitfalls in the PARC model. After all, Xerox failed to capitalize on almost every one of the new technologies that poured forth from the Palo Alto lab. CSL’s aim isn’t just exciting technology, says Tokoro: “The ultimate and unchangeable goal of CSL is to develop technologies that create new markets for Sony.”
Tokoro sees Sony’s future in the emerging ubiquitous network of computers on desktops and in consumer devices. This network, he says, must be regarded as a continually evolving “open system,” a view that is reflected in CSL’s strong emphasis on widely distributed, mobile computing environments where network connections, connected computers and services are changing continuously.
With just 32 researchers in all (including four at a branch lab in Paris that specializes in human cognition), the CSL is a tiny outfit. Nonetheless, CSL’s vision, products and graduates are playing a key role in Sony’s digital transformation. “The CSL is like a pipe,” according to Doi, who is now the lab’s chairman. He counts four laboratories now headed by CSL members or alumni, including his own Digital Creatures Laboratories, which produced the robot dog Aibo. “Today,” Doi laughs, “Sony research is occupied by CSL graduates!”
Unlike the situation at PARC, a steady stream of CSL ideas are now making themselves felt where it counts-in the market. One example: Sony’s ultra-slim VAIO notebook computer is a huge hit in Japan in part because of NaviCam, a key differentiating concept developed at CSL in 1997 by Rekimoto. NaviCam, a tiny built-in digital CCD camera above the notebook’s screen, captures video images that can be edited and transmitted as e-mail over the Internet. NaviCam grew out of Rekimoto’s idea for an advanced type of human-computer interaction in which computers would be aware of people; in development, the concept was transformed into “personal video,” a blend of AV and IT technologies that satisfied Sony’s desire “to promote the computer for more entertainment-oriented uses.”
Perhaps the best example of CSL’s impact on its parent is the lab’s oldest project, the Aperios operating system. After an incubation of six years, Aperios was transferred out of CSL in 1996, together with four or five of its developers, including group leader Akikazu Takeuchi, a 45-year-old former Mitsubishi Electric researcher. Takeuchi now heads a corporate software lab called the Sony Suprastructure Center responsible for home networking and operating system development.
Aperios is a “real-time, object-oriented OS with a reflexive architecture”-a mouthful that means it is particularly good at handling high-speed video and audio streams. It’s this ability that lets Aibo track and intercept the bright pink ball it’s sold with, and also what’s led Sony to place Aperios at the vanguard of the strategy to conquer the home network.
Sony has begun pushing Aperios as the OS for a new generation of TV set-top boxes that are allowing cable companies to deliver hundreds of new channels, as well as interactive services such as Internet access, video-on-demand and games. In May, Sony began selling a set-top box called Plus Media Station in Japan; in October it announced plans to supply New York’s Cablevision Systems with set-tops for its rollout of digital cable to 3.5 million subscribers in and around New York, Boston and Cleveland.
What turns the set-top into the hub of a home network is i.Link, Sony’s version of Apple’s FireWire high-bandwidth interface. The i.Link connection can transmit data at a speedy 400 megabits per second, and Sony already includes it on some camcorders and the VAIO notebook. Last May, Sony teamed with seven other consumer electronics companies, including Philips, in announcing a standard specification-known as home audio-visual interoperability, or HAVi-that should let these companies’ products all talk via i.Link. For consumers, the bottom line on Aperios/HAVi/i.Link is better entertainment: You’ll be able to download the latest Mariah Carey single from the Web via your TV set-top box, route it to your digital video disk player, and be boogeying in no time.
Of course, Bill Gates has his own plans for your living room. Not only is Microsoft promoting Windows CE as an alternative OS for set-tops, but along with Intel, the Redmond empire is pushing its Windows-centric alternative to HAVi called Home API, which can control a home’s lights, thermostat and stereo-from the safety of the PC.
The convergence war isn’t quite as simple as TVs versus PCs, however, since Sony’s vision for your living room doesn’t necessarily rely on a set-top box. “What’s going to be interesting is home networks where devices are capable of advertising their presence and their capabilities and working together cooperatively,” says Rodger Lea, the CSL grad who directs Sony’s Distributed Systems Laboratory in San Jose, Calif., a 70-person outpost of the Suprastructure Center that monitors U.S. developments and supports Sony’s U.S. product groups. “One of the key things we’ve tried to achieve with HAVi is that [like the Internet] it’s very distributed-we don’t want to rely on a single device to be the controller in the home: Any device can be the controller. This is very important for a consumer electronics company which sells a range of appliances. It’s obviously less important for a PC company, which has a very PC-centric view of the world.”
But the situation is evolving almost daily. In the latest twist, Sony and Microsoft have actually joined hands, along with a score of other firms, to advance yet another standard-Universal Plug and Play-based on the protocols that govern the Internet. “What you’re looking at is convergence. What used to be separate turfs are becoming a single market, and there is a huge struggle for who will dominate,” says MIT computer science professor Carl Hewitt, who once taught alongside Tokoro at Keio University and spent time at CSL. “Bar none, the biggest challenge for the CSL is to lead Sony through this paradigm shift.”
It’s a challenge Tokoro is looking toward with confidence. In the future, says Tokoro, “Sony will not only be a leading-edge consumer electronics manufacturer, but also a company which provides network service infrastructure as well as services on the network.” And not just any networked-consumer electronics company, he asserts-the market leader. And his vision for CSL is no less bold. He wants his un-Japanese organization to become the “world’s number one” place for doing computer science research. The next few years will tell whether he’s succeeded, and whether his California-style transplant can lead Sony into the age of the Internet.
The big new idea for making self-driving cars that can go anywhere
The mainstream approach to driverless cars is slow and difficult. These startups think going all-in on AI will get there faster.
Inside Charm Industrial’s big bet on corn stalks for carbon removal
The startup used plant matter and bio-oil to sequester thousands of tons of carbon. The question now is how reliable, scalable, and economical this approach will prove.
The dark secret behind those cute AI-generated animal images
Google Brain has revealed its own image-making AI, called Imagen. But don't expect to see anything that isn't wholesome.
The hype around DeepMind’s new AI model misses what’s actually cool about it
Some worry that the chatter about these tools is doing the whole field a disservice.
Get the latest updates from
MIT Technology Review
Discover special offers, top stories, upcoming events, and more.