Japan tries to race past US; The drive for 'thinking' computers
The Japanese are playing leapfrog - and they are hoping to soar over the back of the US computer industry.
What will put the spring in their jump, the Japanese say, is a $500 million, 10-year plan to develop a ''fifth generation'' computer.
The old cliche is that computers are stupid. Well, that doesn't go for computers with artificial intelligence - a key ingredient in the fifth-generation concept. The Japanese are aiming to produce the hardware and software of a computer that will be able to understand spoken languages, see and comprehend objects, and make decisions based on inference, logic, and deduction. They have formed a separate institute, sponsored by the government, to research such a computer. Eight leading Japanese makers of electronic products, in addition to government-controlled Nippon Telegraph & Telephone, will also be involved in the project.
The strategy of the Japanese is that success in the program will catapult their industry over the bulk of computer research now going on in the United States - and over Japan's sad performance in the worldwide information industry. Ten to 15 years from now, when the project should be finished, Japan expects to emerge as the industry leader.
''Their written plan clearly states 'world supremacy,' . . . but I don't think the battle is lost,'' says Prof. Michael Dertouzos, speaking of the ability of the US computer industry to meet this challenge. Dr. Dertouzos is director of the computer science laboratory at the Massachusetts Institute of Technology, a lab that has devoted considerable effort to the fifth-generation field.
The US is now taking a few tentative steps to ready itself for this and other Japanese thrusts at American industry:
* In mid-August, the Microelectronics and Computer Technology Corporation (MCC) became a legal entity. This corporation, launched by Control Data Corporation, will be a consortium of US semiconductor and computer manufacturers. The companies will invest funds in MCC, and the funds will be used for research and development (R&D) in areas of key interest to the industry. Digital, Burroughs, Sperry, Westinghouse, Xerox, Motorola, and about eight other firms are now considering a commitment to MCC.
* In May 1981, the Semiconductor Research Corporation began its first fiscal year. This group is a consortium of companies which use, sell, or make semiconductors. Participating firms in SRC provide funding to universities on semiconductor research projects. One aim of SRC is to improve the university education of engineers - which the industry now complains about. Both consortiums represent the first attempts by industry members to cooperate on R&D.
* The Department of Defense (DOD) - virtually the only active government subsidizer of R&D in the computer and semiconductor industries - wants to increase its funding for high technology projects. Richard D. DeLauer, undersecretary of defense for research and technology, was quoted recently in Aviation Week as being particularly eager for the department to spend more on artificial intelligence research.
''We are in an excellent position to get there first, but we can't get our act together,'' says Stanford University Prof. Edward Feigenbaum. Last fall he was invited by Japan to lecture at its international conference on the fifth-generation computer.
Japan's act is a tough one to follow, not only in their fifth-generation plan , but in the high-tech industry in general.
Ira Magaziner, a corporate-strategy consultant and author of ''Research and Development in Japanese Industry,'' ticks off advantages to the Japanese system that ''in the next decade will put the same pressure on (US) elctronics industries and users as we've seen in past decade on cars and steel'':
Education. Japan has an active training program for its personnel. Twenty to 40 percent of a company's engineers are abroad studying every year, Mr. Magaziner says.
Long-term research and development incentives. Japanese companies get substantial tax credits, grants, and loans from the government and banks to help fund R&D. The cost of R&D in Japan is not a great concern to companies, says Magaziner, because of government incentives, and because firms focus on revenues and not immediate profit. Also, firms often share basic research.
Government tax breaks to companies that use evolving prototypes. The result is a ready market for new products.
US industry is aware of ''how beautifully,'' as Stanford's Professor Feigenbaum put it, the Japanese system works. The question is, how to compete with it.
Most consultants, computer scientists, and industry leaders in the US agree long-term R&D efforts need to play a greater role here. But with the US emphasis on short-term profit, it's wondered whether such a change can be brought about.
''We should not worry so much about how much money we can make next month. Instead, we should concentrate on how we can do a better job next month and every month thereafter - a better job of productivity, a better new product, and better quality,'' ran a statement by David Packard, chairman of the board at Hewlett-Packard Company.
But William J. Weisz, vice-chairman of the board at Motorola Inc., said in a phone interview: ''Industry has changed. We have been doing a superb job in better service and quality. When prices (of semiconductors, for example) come down at ridiculous paces, it puts an awful lot of pressure on American companies who, unlike Japanese firms, have to fund their own growth and own research and stay afloat in the present financial environment.''
''We can't mirror Japan,'' agrees William Shaffer, a spokesman for the MCC consortium. ''We can't do it because in this country we don't bureaucratize creativity. But we can learn from Japan.''