The Future of Microcomputing

---

The Future of Microcomputing

	What will the standard PC offer in five years?	What new features will come in 10 years?	Will tomorrow's computers be predominately general-purpose or specialty (i.e., Web PC) systems?	How will microprocessors change our homes in the next decade?	What kind of new GUIs will more powerful processors make possible in 10 years?	How important will the Internet be in five years?	How long will Moore's Law continue to be relevant?	Whe n will quantum effects and other problems require radically new chip technologies?	Will we manufacture chips in zero-gravity environments in orbital fabs anytime soon?	What will be the next "sea change" in computing?
Marc Andreessen, Netscape Communications	It will have a thin, rich-color screen; process spoken commands; integrate voice, video, and screen-sharing with powerful visualization tools; run about 10 times as fast as today.	Another factor of 10 in performance and storage capacity. People will wear computers tied to lightweight "mirror shades"-style displays.	Specialty systems may outnumber PCs in five years, but even more important will be ubiquitous computers for flexible, secure access to your information, wherever you go.	At least one-third of American homes will integrate entertainment systems and satellite, cable, Internet, energy, and l ighting technologies with networked information by the end of 10 years.	The idea of a GUI is too restrictive. Imagine an immersive user interface that leverages global, real-time, multimedia, and networked information.	The Internet will have begun to disappear, like electricity and telephony, into the woodwork. And it will [be accessible] everywhere, in wired or wireless forms.	Through 2020, when we will see a discontinuous improvement in performance rejoining a new Moore's Law curve based on a transition toward molecular nanotechnology.	Probably until about 2007. By that time, however, dramatically new chip technologies based on quantum dots and tunneling will have begun to arrive.	Only as an unlikely and distant possibility. Nanotechnology will have begun to bear fruit before zero-gravity chip manufacturing makes sense.	A secure, truly mobile agent language -- way beyond Java -- will eliminate the Tower of Babel that prevents us from harvesting more of the benefits of computing and communications today.
Dan Dobberpuhl, Digital Equipment Corp.	Eight times faster per CPU (with symmetric multiprocessor options), four times more memory, flat-panel touch displays, speech recognition, eye tracking, and videoconferencing.	Forty times faster per CPU (four-way SMP, gain depends on software), 16 times more memory, and [who knows what else?].	They will coexist. Over time, specialty devices will dominate because they'll be cheaper. General-purpose devices will remain important, especially in business and technical applications.	Embedded microcontrollers are already pervasive. In 10 years, they'll be linked by wireless networking within the home, leading to all sorts of wild possibilities.	There are only so many things to be done with GUIs. The real advances will be i n how the computer gets its inputs from humans, with things like speech recognition, eye tracking, etc.	Demand for high-performance internetworking will continue to outstrip the ability of the communications infrastructure to actually supply it.	For at least 10 years. The rates of improvement are likely to drop off somewhat, however, due primarily to economic considerations rather than technical limitations.	Quantum effects are here today and must be taken into account. But they won't force fundamental changes within the 10-year horizon.	No.	The merging of cellular phones, portable computers, and high-speed networked servers offers many exciting possibilities.
David Chaum, DigiCash	This is sort of like asking, in 1896, what the office typewriter will look like in 1996.	[See #1]	The location of disk drives is n ot the issue. But seamless downloading and full associative linking of software and information, together with a suitable payment method, holds enormous promise.	Integrated systems will access text, voice and video messaging, interactive entertainment, and electronic commerce. But let's hope that our children still learn to program!	Many people will be able to communicate verbally with a portable digital "representative": an on-line business agent and personal secretary. Some representatives will be at home in 3-D virtual worlds.	At least as important as the combination of phones, TV, radio, printed publications, and PCs are today.	The enormous investments already made should allow current chip technology to continue its price/performance trend for five to 10 more years.	I'm not sure we will be forced to develop nanotechnology, but I sure hope we do.	There's some future for it, but the extent of i t remains to be seen.	We may experience a gradual drift into a surveillance society. Alternately, cryptography and digital "representatives" will protect our privacy while allowing participation in cyberspace.
Andrew S. Grove, Intel	Multiple video windows for desktop conferencing, intelligent agents for filtering and finding information on intranets and the Internet, rich 3-D graphics. All managed remotely.	Any answer, from anyone, is science fiction.	The PC will still be the mainstream machine, but there will be a wide range of specialty peripherals that connect to it.	In five years, you'll play games on a PC instead of a TV, do homework on a PC instead of on paper, and communicate via your home PC.	Ever-increasing processor performance enables ever more photorealistic and 3-D GUIs. I expect that this trend will continue.	Huma ns subtract out powerful mediums -- such as radio and TV -- from their experience and focus on the content. The Internet will be the same way, vastly important, and just there.	It will be stable -- for at least the next 15 years.	Again, for at least the next 15 years.	No.	If I knew that...
Doug Engelbart, Bootstrap Institute	The standard PC may well be your personal PC rather than the office PC. It will probably be a descendant of the network computer that we're talking about today.	Surgically implanted?	The network computer is going to be the general computer. The idea of locking up all your resources in one box will disappear.	Home-based education, communications, collaboration, and telecommuting will expand. Appliances, lighting, and other things in our homes will become more flexible.	GU Is are metaphors of yesterday's way of living -- a desktop, a folder, and so on. We need to think of ways to connect humans to [information] that are not related to the underlying system.	Eleven on a scale of 10. That kind of communication is priceless and inevitable. We absolutely have to think of the future with a public, global, high-speed network.	The limit for miniaturization, speed, and low cost [is] beyond where semiconductors can go. In 10 years, another technology might replace semiconductors.	We'll run out of gas in 10 years. Eventually we'll be arranging individual molecules.	Zero-gravity is great for some things, but what? On many processes, gravity has negligible effect. Surface tension, for example, can be a greater factor.	Improving the collective IQ for people who want to collectively work on tough problems. That's the grand challenge. The first one to do it wins, and the real winner is humanity .
Federico Faggin, Synaptics	Eight times more RAM, eight times more magnetic storage, four times faster processor speed. Built-in support for teleconferencing and intelligent databases.	Seamless integration of computing and communication, pattern recognition for speech, and simultaneous language translation for simple sentences.	The general-purpose PC will continue to be dominant for the foreseeable future, but there is room for specialty systems.	Our homes will enjoy on-demand entertainment, virtual reality games, and a critical mass of multimedia and interactive education software with the associated social impact.	We'll manipulate simulated 3-D objects and navigate in a 3-D cyberworld. The computer will have the senses of touch, hearing, and sight, making possible human-like interactions.	It will have major social impact and vast technological and economic consequences. It will promote peace by bringing people together and will restructure the communications infrastructure.	For at least another 30 years. Moore's Law will progressively slow down , from doubling every one-and-a-half years now to doubling every two years starting in five years, and so on.	A factor of 1000 in circuit density; a factor of 100 in chip area. From one active layer to multiple active layers, up to 100 layers; overall a factor of 1,000,000.	While research may be important, I see insignificant manufacturing in space for at least the next 20 years.	Computers with sensory-motor capabilities that turn into autonomous, intelligent machines capable of purposeful and intelligent behavior in the real world.
Jerry Rogers, Cyrix	The office PC will have a speech-enabled human interface with the ability to access, process, and store informati on all over the world.	The desktop PC will disappear, replaced by a pocket-size mobile device with 10 times the power of contemporary PCs.	In five years, the vast majority of computers will be general-purpose PCs. In 10 years, the volume computer will be a hand-held device for wireless, e-mail, and Internet communications.	In five years, homes will have a central PC serving as the communications backbone for the family. In 10 years, there will be a major shift to personal mobile communications and computing.	Dramatic improvements in microprocessor performance will finally bring computing to a new human-interface era of speech and virtual reality interfaces.	The Internet will become the primary source of information, training, and customer support both at home and at work.	Three-dimensional imaging and rendering, speech, Internet agents, and the human interface will drive the almost insatiable desi re for computing power indefinitely.	The challenge is to drive on-chip voltages to 1 V as we approach 0.1-micron gate geometry to avoid quantum effects. This trend will probably continue for 15 years before we hit the brick wall.	Capital costs of space manufacturing are and will remain prohibitive for the foreseeable future. It's also hard to find people with both manufacturing and astronautical skills.	The human interface will finally fulfill the ease-of-use model for everyone to feel comfortable. The interface will consist of sophisticated speech recognition and synthesis and virtual reality.
W. J. "Jerry" Sanders III, AMD	It will likely operate in the range of 300 to 700 MHz. Its performance will scale linearly with frequency, so a 500-MHz PC will be five times as fast as today's 100-MHz PCs.	The networking infrastructure will begin to catch up to the capability of personal computers. Currently, this infrastructure hobbles their capability.	In five years, there will be some fragmentation toward specialty devices. But because this is a cost-driven industry, general-purpose computers will dominate.	Fat pipes will make our phones and TVs more PC-like. Ever more powerful microprocessors will enable computers to become a central communications tool or port in our homes.	Untrained speech and handwriting recognition will be commonplace. Motion picture-quality video and 3-D graphics with surround-sound will be directed by voice command.	The Internet will be as ubiquitous in our lives as cable television is today.	While it becomes increasingly expensive and the slope of the line may flatten a bit, I don't see an end to the progression over the next 10 years.	Silicon has legs. We will still be using it below .1 micron!	Truly a dumb idea. In our c ost-driven industry, there would not be an acceptable economic return.	Fat pipes. The fifth wave of computing, i.e., public-network computing, will realize its full potential when bandwidth is as cheap as bits.