June 2, 2000
K. J. Kearney
R. N. Seitz
The future of 2020 will be many things, but the best word to describe the future will be: interesting. It will be far from the dystopian, robot-reigned world of certain science fiction authors, with computers being more helpful than competitive.
There may be some hiccups on the way to chilling visions of robotic domination. For one thing, it isn't entirely certain that we can devise robots more intelligent than ourselves. It isn't a given that computer hardware will become fast enough and cheap enough to support the hardware requirements for human-level intelligence before we hit a technological wall. (My personal bias is that hardware will become sufficiently cheap.)
We don't need to make them smarter than ourselves. There are people among us who are 2.5+ times as smart as the average bear, and yet, we don't worry about their conquering the world and enslaving or eliminating humanity. We've been worrying about artificial intelligence at least since Mary Shelley wrote "Frankenstein".
We anticipate computer augmentation of the human mind in ever more intimate ways. By 2020, we might see some kinds of more-or-less direct couplings between computers and the brain. These might take the form of sensors that could detect subtle motions of some lesser muscles, hearing-aid-type earphones, and glasses upon which computer images are superimposed. Implants have been forecast for 2030. We could imagine a future in which we draw upon computers for support while choosing to retain our comfortable ways. We would hope to see artificial intelligence used to amplify human intelligence rather than replace it.
For instance, you could be awakened by the music coming out of your computer, which, during the night, has searched the internet for songs that you would like based on your personal preferences and downloaded them. Your computer has also printed out today’s paper for you to read over breakfast. Much better than going out in your bathrobe and removing your paper from the bushes. Alternatively, newspapers will probably be video enhanced, digital sound enabled and hyperlinked so that your favorite family member or celebrity might read you the morning news and be prepared to discuss the issues with you real-time.
You walk into your kitchen, making sure to play with your RoboDog on the way there. When you reach the kitchen, your refrigerator’s LCD screen informs you that you are out of both milk and eggs, and offers to call the grocery store for you. Oh well, there goes breakfast, you think to yourself. You look outside to see your AutoMower driving carefully around your rose bushes on its automated track. You sit on your couch and call out “Television On” which slides a panel back on your wall to reveal your TV, which has already recorded your favorite shows from last night and deleted all the commercials from them. Alternately commercials might be the preffered viewing experience and they will be recorded, categorized and hyperlinked to enable online purchases or purchase research from your living room couch.
You call up your e-mail account and read that someone has hired your services as a mercernary on your favorite online game. They have already transferred your fee to your bank account. Maybe that’ll help you pay off your medical bills. (You’ve always had 20/20 vision, but you just had to get that new laser operation that promised “hawk-like vision”. Comes in real handy at football games. Now you don’t mind the nosebleed seats as much.)
The world of 1980 wasn’t terribly different
from the world of 2000, and I think the world of 2000, looking back from
2020, won't seem that much different from the world of 2020.
Smart Highways, Vehicles
Intervehicle, Highway Intercommunication
I'm surprised that we aren't already utilizing intelligent, "ground-traffic-control" highways upon which vehicles know their exact positions and orientations, and therefore, their speeds. Such a highway system could monitor the positions and velocities of all its tenant vehicles, and would be able to provide automatic lane signaling, braking, steering, collision warning, and collision avoidance, including emergency takeover in case the driver became incapacitated. The highway could choreograph the motions of all its resident vehicles. Wires buried in the roads, carrying weak RF signals, magnets embedded in the roads, or some other RF grid-work could allow vehicles to be located within, perhaps, a foot.
Highway plans don't seem to be headed in this direction, so it remains to be seen whether or not autonomous vehicle control will follow this approach. It seems probable that each vehicle will be computer enhanced with an artificial intelligence to take over routine driving chores. They will probably operate networked, passing along road information transparently as packets of information as they pass each other.
Computer-assisted driving and flying:
This is another approach to autonomous vehicles and the one that appears to be the front-runner at the present time. A road-following, machine-vision-activated turn signal is now appearing in Mercedes trucks, as a trial-balloon for ultimate road-following visual navigation systems for trucks. These features will debut first in high-end automobiles, gradually diffusing downward to economy cars and trucks. Automated road accident systems will prevent traffic piling up behind disabled cars and trucks.
Airplanes will be able to check on their pilots, crew and pasengers and radio for help/instructions in the event of emergency or disturbance. Aircraft flying in the same airspace will “know” where other aircraft are for collision avoidance and ttraffic control by using a real-time satellite network.
By 2020, petroleum prices should be soaring. Also, anthropically-generated CO2 isn't going to be very popular. I think a momentous change in our automotive habits is in the offing.
Fuel Cell Powered?
One problem with fuel cells is that of avoiding fossil-fuel consumption in generating their fuels. However, hydrogen/oxygen fuel cells presently seem favored among alternative vehicle propulsion schemes.
Alcohol-powered vehicles were used in Germany during World War II. Farmland is required to produce alcohol, but alcohol-powered cars and trucks would utilize renewable fuels, and wouldn't add to the net CO2 content of the atmosphere.
At some point in the future, the current Moore's Law 100-folding of transistor densities is going to break down. Progress using a given technology tends to follow an S-curve. Right now, we're still on the linear portion of the sigmoid curve. Intel, TI, and IBM have announced plans for conversion to 0.13-micron geometries at the end of this year and early next year, and a transition to 0.10-micron geometries late in 2002, keeping us on the Moore's Law curve until early 2004. Further migrations to 0.07-micron geometiry and then to 0.05-micron geometiry appears feasible and likely, carrying us through late 2005. I'm going to suppose that we might remain on the Moore's-Law curve more or less through 2010. However, remaining on the Moore's-Law curve through 2020 would bring us down to 20 nanometer or 6-atom-wide features. This might or might not be possible, so, for the lower limit on 2020 design features and capabilities, I'm going to use the 2015 Moore's-Law projection, and for the upper limit on 2020 circuit features and capabilities, to utilize the 2020 Moore's-Law projection. It's worth noting that even if we "hit the wall" in computer price-performance improvements, prices will continue to fall for 5 to 6 years after this happens because this is the time required for prices of cutting-edge computer devices to bottom out.
I believe that we will reach the levels of computer price-performance values that will allow human-caliber artificial intelligence hardware.
2020 Computer speeds:
1 to 10 terops
2020 RAM Prices: $1,000 to $100/terabyte RAM
2020 Hard Drives: 4 to 100 terabytes of disk
2020 Laser-Optical: 100 to 1,000 gigabyte DVD’s?
2020 Internet Bandwidths: 10 to 1,000 megabaud wideband
2020 Webcam: Yes
2020 Displays: At least 2,000 by 1,500-pixel HDTV displays (see below).
2020 Digital Signal Pr.: 10 to 100 terops DSP’s
These numbers reflect 1,000-fold improvements (per Moore's Law for 2015) to 10,000-fold improvements (per Moore's Law for 2020) in computer speeds and storage capacities over their current Year-2000 values.
Embedded computers (in trucks, washing machines, TV's, etc.) are probably more important than desktop computers. However, the above numbers should also be indicative of the performances that might be expected in embedded computers.
I think the most significant thing about robotics is that it's already very much with us. Fifty years ago, there was a great deal of concern about robots taking over. We would face massive technological unemployment. Common folk would become a bread-and-circuses proletariat, on the dole, as they were replaced by robots.
In Karel Capek's 1926 play "R. U. R". ("Rossam's Universal Robots"), the robots take over, eliminating humanity. In Jack Williamson's "The Humanoids", the robots lobotomize humans to ensure their contentment, and to protect them from harm. By now, we've had fifty years of automation, including Puma and Unimation industrial manipulators, and it hasn't happened yet. Telephone operators have been replaced by crossbar switches. Clerks have been replaced by computers. Clotheslines have been replaced by automatic clothes driers. Assembly-line workers have been replaced by specialized machinery. Typists have been augmented by word processors, with voice dictation in the offing. Manual car-washing has been superceded by robotic car washers. Full-service gas stations have given way to do-it-yourself gas stations. Mom-and-pop grocery stores have been exchanged for supermarkets. And yet, we have an unemployment rate below 5%.
A few weeks ago, our trash service was automated. Where three men performed our garbage pickup, only one is now required. (The other two men on each truck have been reassigned to other jobs.) The mechanism that picks up the trash consists of two hydraulic arms that sense the trash can, embrace the trash can, lift it high over the truck, and shake its contents into the truck's monstrous maw. The trash mechanism fits all the requirements for a true robotic device. (I'm just waiting for a James Bond movie in which 007 hides from the bad guys in a trash can and is dumped into the trash truck.) Gradually, we have become more and more automated, but slowly enough that we hardly notice it.
I would expect to see a continuation of this trend through 2020. The main market for robotics will probably lie in the commercial world, as it has since water-powered Jacquard looms evoked riots in the streets of Paris by the displaced weavers.
I could imagine some anthropomorphic robots by 2020, but I wouldn't expect every household to own one. Maybe by 2030...
Visually-navigated autosweepers, autoscrubbers, automowers. Yard trimmer/edgers.
I expect to see the first visually-guided automatic lawnmowers and carpet sweepers to make their commercial debut in 2005. By 2010, I think some lawn and cleaning services may be using them. Sometime between 2010 and 2015, I think we might buy them when they go on sale at Walmart.
By 2020, I think lawn services might be getting some competition from automowers/trimmers/edgers. These devices will be nothing more than automatic machines, like automatic washers and dryers.
I could imagine inflatable human surrogates, with lightweight "bones" and actuators, that could carry on very limited conversations. We'll probably see such Animatronic figures at Disney World. I could also imagine cyber-companion computer programs that could interact in a limited way with their human owners. Automated mail carts, serving carts, personal living assistants for the disabled are all probable. Automated tennis and golf pros should become a reality.
Twenty years ago, LCD's were limited to tiny black-and-white calculator displays. The CRT was the only game in town. Ten years ago, there was a lot of interest in plasma, micro-mirror, ferro-electric, cold cathode, and electroluminescent displays, and thin-film transistor (TFT) liquid crystal displays.
Today, the TFT LCD, plasma, and micro-mirror
displays are commercially available, although the CRT is still the leading
display contender. Projection displays may have improved in quality, but
they haven't dropped much in price. The message is that changes in display
technologies, as is the case in other technical domains, are slow to catch
on. One new sleeper in this contest could be organic displays. These offer
the promise of being producible by ink-jet printers that would spray a
conducting grid on a plastic substrate, covered by organic chromophores.
Nanotechnology displays are also a possibility. Over a 20-year period, there might well be a major shift in display schemes, although the venerable CRT might still be alive and well 20 years hence. Holographic dispays are now in use and might become widely available, as well as heads up displays, and retinal projection.
Motion-Picture Experts Group 4 (MP4) compression, in its fullest flower,
animates 3-D "shell" models in a receiver, using animation tracks and 3-D
still imagery transmitted over a communications link. Rather than send
essentially the same image 30 times a second, it downloads three-dimensional
source images or "texture maps" of objects at the transmitter to a computer
at the receiving end, followed by animation instructions. The downstream
computer, using the imagery it has received from the transmitter, reconstructs
the images of these 3-D source objects and animates them in accordance
with the animation instructions it has also received from the transmitter.
This can provide a huge reduction in bandwidth. (Note that news broadcasts
in which panoramic vistas are presented and in which scenes may change
fairly rapidly wouldn't profit much from MP4 compression.) By 2020, MPEG4
may be widely used for the recording and display of imagery.
It has been said many times that the Internet will become invisible and taken for granted. It will probably become our primary information network. The key to its capabilities is bandwidth. I would hope that we are operating in the tens of megabaud range by 2020. That will require major expansion of Internet backbone capacities, but they're expanding by leaps and bounds already, now that the majority of homes in the U. S. are on the net. It will require better linkages into and throughout the home, but 2.3 GHz RF links could economically provide that capability in existing homes at low cost without rewiring.
Alternatively, rewiring of existing homes with fiber-optic cable shouldn't necessitate much effort when existing telephone cable is used to pull through the fiber-optic cable that replaces it. The costs of fiber-optic hardware for the home are dropping rapidly. Right now, the missing element is the link from the neighborhood telephone projectors to the curbsides of neighborhood homes.
It will be a few years before there is heavy penetration of medium-band access to the world's homes. However, by 2020, we should enjoy much faster access than we now experience. Given wideband access, a number of changes might be in store for us.
The downloading of MP4 video may follow the current downloading of MP3 music.
Children (and adults) may play video games
over the Internet (as they do today), but with much more realistic involvement.
Surely by 2020, virtual reality will become a major entertainment modality.
E-commerce will be bigger and better than ever by 2020. Virtual-reality shopping experiences, in which the prospective buyer can look at the imagery of objects from various angles, or perhaps, see it live in local stores via cheap overhead webcams, might become popular.
One problem that should be solved by that time is the economical billing of small amounts (e. g., $0.25) on credit cards. I doubt that automated delivery trucks are on city streets by 2020, but if and when that arrives, it might lower the cost of delivery services.
With the decline of American public schools and the rise of home-schooling, there may be a major market for distance learning at the primary and secondary school level, perhaps to supplement existing programs. (Alternatively, there may be computer-based teaching in the schools themselves.)
There would seem to be a market for this among the gifted and talented. Our present lock-stepped, one-size-fits-all system of graded education is better than no education at all, but it doesn't fit anyone terribly well. The presumed wideband access of 2020 might render this an attractive option. It can be made interactive, and can draw upon the best course material and the best teachers in the country.
A market for distance learning certainly exists for university-level training. Here, cost, convenience, and accessibility are the drivers. It seems to me that U. S. universities are pricing themselves out of the market. Also, it can be next-to-impossible for someone who's traveling on the job to complete courses in a conventional classroom setting. On-campus exposure may be a valuable experience when it can be done, but in many cases, it simply isn't feasible. Here is another area where immersive environments and virtual reality experiences might be very useful.
At the same time, there are fields such as medicine or dentistry requiring hands-on training in which distance learning can't be used. And it can be valuable to experience a classroom environment and to interact with faculty members and other students. Teenagers may profit from getting away from home, and out on their own.
Computer Assisted Classroom Teaching
Perhaps by 2020, there will be a computer for every student. Alternatively, given super-cheap displays, one might use terminals at every desk. Free high-quality software would also help.
I think schools are here to stay, but distance learning and "canned" learning may supplement conventional personal-contact programs. Distance learning might even complement on-campus labs and hands-on experience. Distance-learning programs could be accessed day or night (unless they're "live"). This could even be useful in college dormitories.
Besides being tailored for the individual child or adult the learning programs will be hyperlinked and mult-media to enable the user to customize the overall result using the generic lesson plan as a template for a personal learning experience. Degree programs will more resemble a mosaic than the current patchwork quilt.
Search engines should be out-of-this-world-good by 2020. MPEG6 will allow someone to hum a few bars of a tune, which the search engine will then locate. Search engines may become smart and interactive, with dialogs with the user to narrow the field of search.
The SETI (Search for Extraterrestial Intelligence) program has demonstrated how personal computers can be used in a background mode to solve problems of general interest. As bandwidths rise, personal computers could be networked to provide a supercomputer capable of solving massive problems that lend themselves to a degree of compartmentalization. This might conceivably be used for artificial intelligence (AI) experiments in which the AI's experiential database is distributed across the network. It's also possible that robots could be linked via wideband RF to the Internet, with object recognition and cognitive functions performed by a remotely located computer or computers. In principle, the entire Internet or some sizable subset of it could become one big supercomputer, limited only by bandwidth constraints and by the ability to perform cognitive functions in parallel.
Internet-based supercomputers, AI
If you want to worry about artificial intelligence taking over the world. this might be a good place to start. Sandia Corporation is developing a "web-bot" that will sniff out attempts to hack government computers and to launch cyberattacks upon them. Sandia isn't worried about amateur hackers but about hostile governments that might want to bring down our military, and other federal, computers in a concerted attack. One of the concerns is that a foreign power might conscript many Internet computers and blend them to form a supercomputer. The Sandia "web-bots" would reside on computers on the Internet and would be decentralized so that they couldn't be disarmed from a central site. They will be delivered to government installations next year, with the possibility of commercial availability in 2002 or 2003. However, the idea of these software entities having a life of their own and being difficult to eradicate is a little creepy in itself.
Could anything go wrong? How wrong? Could a supreme artificial intelligence develop on the Internet? Could it commandeer other computers to enhance its intellect?
By 2020, there may have to be restrictions regarding what can be published on the Internet, with particular attention to intelligent agents.
Agents will be in use to organize and dispaly useful internet information that is customized to your circumstances, bringing diversion or supplemental information or personal networking as required, on demand, real time. Intelligent agents should actually be able to generate new internet content as required by 2020. They should be able to auto-generate documentaries, catelogs, pamphlets and reports on command.
By 2020, I would expect film cameras to be restricted to highly specialized applications, if not relegated like slide rules to museums.
Digital cameras might transmit their pictures over the wideband cellular links to free storage sites on the Internet, or to one's home computer. Cheap webcams will be everywhere. They're already down to the $25 to $35 price range. Webcams might transmit their video imagery over wideband cellular links to remote storage sites. Cameras may actually become virtual with image recording being an integral part of several devices like your car or phone with access on command as required.
There will probably not be any difference between webcams and video cameras, or between video cameras and still cameras. (One camera does it all.)
I would also expect that we would have spherical, machine-vision "eyeball" cameras about an inch in diameter that would either have dual sets of CCD's or CMOS sensors, or would have a graduated-resolution imager like the retina of the eye. This camera would have autofocus and a colored auto-iris, and would look like a human eye.
The difference between animation and filmmaking will continue to diminish, with actors and directors, backgrounds, and sets being generated within the camera system. Movies will no longer be viewed as a series of static events that can only unwind one way. Movies and probably television should evolve as a viewing arcade with multiple plot threads at varying levels of detail available simultaneously on command.
I would predict a heavy emphasis upon virtual reality, including, perhaps, cybergloves and motion seats by 2020. An attachment called "iSmell" will be available by Christmas of 2000 that uses synthetic scent cartridges to provide fan-wafted odors on command. It will cost about $50 in the low-end version. Will it sell? We'll see. Holographic projections will probably be the norm.
Simulations could also be a good video game/learning tool. One might explore an interactive cyber-zoo with strange alien animals, or could visit other countries.
Learning adventures would be another desirable kind of video game. Artificial intelligence may be a staple of 2020 video games.
Very effective “smart pills” should be available by 2020. A Johns Hopkins spokeperson has advised her audience that one or more of the Alzheimer's drugs currently in the FDA pipeline should boost children's IQ's by as much as 50 points by 2010. One pair of researchers is testing a pharmaceutical that could give people total long-term retention for a few hours. This agent should be available in 2 to 5 years, although it will probably be available by prescription only. One would have to take care that one didn't memorize license plates or other trivia during the 3 or 4 hours the drug is active. These agents could revolutionize the world, and would afford the advantage of controllability. (At the same time, tests like the SAT and the GRE would become a joke.) Given these pharmaceuticals, humanity could boost its own intelligence. Of course, it would take a while to determine the long-term safety and side effects of such treatments.
The development of "Doogie Hauser" super-smart mice suggests that such gentic manipulations are possible. There is at least one gene that codes for intelligence.
By 2020, I would expect that humanity will have the ability to produce genetically-smart animals, and genetically-smart people. How we will choose to use that is open to question, but the possibility of boosting chimpanzees to near-human intelligence might reveal itself. In reality, this will be a hot research topic, but it will undoubtedly be scrutinized to avoid dangerous experiments. In the meantime, humanity will have a means of boosting its own intelligence considerably.
By 2020, children's toys may be expected to contain microprocessors that are as powerful as our fastest supercomputers. This may lead to conversational toys that can carry on rudimentary conversations with children. Such microprocessors might well afford primate intelligence, and possibly, a watered-down version of human intelligence. (Can't you imagine the scare movies?)
Robotic pets and robotics kits may be other popular options for children. Portable, transparent, low-cost computers and/or Internet appliances may be available for children. Most children may carry low-cost cellular phones so their mothers can reach them. Robotic companions may be the avenue for computer assisted teaching, with the programmed mission to monitor performance, encourage achievement and customize learning experiences as required to insure lessons are truly learned.
I would expect to see capable artificial eyes by 2020. Alternatively, the regrowth of one's own organs from one's own cells might be in the cards by then. It would be wonderful if we could regrow coronary arteries, and hip and knee joints.
In the area of genetics, it will be the most significant genes that will be tackled first. Major diseases will be targeted, along with intelligence and longevity. (Intelligence- and longevity-boosting are enabling improvements that can make everything work better. It's like getting three wishes and using the first wish to wish for three more wishes.) I would be hard-pressed to come up with a timetable for genetic discoveries because they are srendipitous. However, cancer, cardiovascular disease, and Alzheimer's disease will be bellweather challenges.
The dark side of biotechnology is biological warfare. Another dangerous aspect could be hackers who could create insidious lifeforms.
Biotechnology may point the way toward
replacements for antibiotics.