The Harrow Technology Report
http://www.TheHarrowGroup.com
Insight,
analysis, and commentary on the
innovations and trends of contemporary computing,
and on its growing number of related technologies.
An ongoing journey towards understanding,
and profiting from, a world of exponential
technological growth!
Copyright © 2001-2005, Jeffrey R. Harrow. All rights
reserved.
Email: Jeff@TheHarrowGroup.com
YOUR
Views of the Next 'Killer Apps.'
A Special Report
Adjunct to this week's
issue, at
http://www.theharrowgroup.com/articles/20021118/20021118.htm
Nov. 18, 2002
In a recent issue (http://www.theharrowgroup.com/articles/20021104/20021104.htm#_Toc23925811)
we explored how, in about eight years, commodity
CPU chips may be running at 15,000 megahertz (15
gigahertz) and be powered by one-billion
transistors. That's some SERIOUS commodity
computing power.
Considering that in Nov, 2002
the latest and greatest will be performing at
3,060 megahertz (3.06 gigahertz), and that most of
us would be very hard-pressed to use all of that
power (excluding scientific and power users, and
gamers), I suggested that what we REALLY need to
rekindle the "PC Revolution" is a set of new
"Killer Apps;" their features will be SO
compelling that, once again, we'll all be on the
edges of our seats waiting for the next notes of
Moore's Law to hit the shelves.
But what will those Killer
Apps be? And so far, the software industry
hasn't come up with a new one for quite some time.
So I asked YOU what
applications YOU think might again light up the PC
world, so that your ideas might spur software
developers in these direction. Your
responses have been, well, 'bountiful' to say the
least, and I'm going to share a selection of them
with you here.
I've taken the liberty of
trying to make sense of this technological
future-brouhaha by consolidating them roughly into
sections. But as you'll see, many ideas want
to span a number of areas.
May you, and the developers
who read this, succeed in again lighting up our
"PC lives."
Click on either
the Categories, or on the individual
suggestion links, in the Table of Contents below:
Virtual Reality.
-
Virtual Reality, by Miriam English --
-
VR, The 'CPU Killer,' by Scott Miller --
-
VR Entertainment, by Des Markland --
-
Immersive Entertainment, by Jim Van Vorst --
-
Virtual Reality, by Thor --
-
Two Words: Virtual & Reality, by Derek Mathias --
-
"Altered Reality" Videoconferencing, by Kirk
Hutchinson --
-
Fully-Immersive VR, by John M. --
-
"Usability" Is The Key, by Nick Gassman --
-
Holographic Displays, by Wil Marshman --
-
"Holo-CAD Drawing", by M. Williams --
-
3-D Output; Virtual and Real, by Geoff Keller --
Extending Reality:
-
Bandwidth - The Enabler, by M. Richter --
-
Bandwidth!, by John Matrow --
-
Video Conferencing, by Tom Williams --
-
Video Instant Messaging, by Mark McCoskey --
-
Virtual Visiting, by Derrick Davis --
-
Virtual Tours, and More, by Byron Law --
-
The In(puts) and the Out(puts), by Bart Wessel --
-
"Tactile," by Dr. Ed Reifman --
-
PC As Phone, by Martin Pensak --
-
Embedded Full-speed Video, by Bobski Masson --
My Life, My Bits:
Digital (Two-Way) Voice:
-
True Voice Response, by Richard Dinning --
-
The Butler, by John Smith --
-
The "Other" Butler, by Max Rible --
-
The Butler, Times-3, by Kim Mains --
-
Computers That Listen, by Jeff Allen -
-
"SpeakEasy," by Rocky Rawstern --
-
Communicating Verbally, by Andrew Clark --
-
Superb, And "Learning," Speech Recognition, by
Jack Lipscomb --
-
Making Computers "Disappear," by Jeff A.K. --
-
Voice Recognition Will Enable MANY Apps, by Ronald
Kaledas --
-
REALLY Effective Voice Recognition, by Adrian
Salone --
-
"Listening," But to the BRAIN (and more), by Shawn
Gold --
-
The Invisible User Interface, and "Vertical
Intelligence," by Henry Nash --
-
Personal Robots, by Martin Spencer --
-
Video Editing Power-Up, by Mel Lammers --
Interface, Usability, and Moore:
-
The More Things Change, The More They Stay The
Same, by Chris Denver --
-
"Interface," by Cameron Reilly --
-
The 'Big Book,' by Michael Siwinski --
-
Home Entertainment Center, by Joel Millett --
-
Complexity - Of Movies, Liberty and More, by Grant
Perkins --
-
Remote Medicine, by Joe Laberge --
-
Video Recognition, by Michael Wiedower --
-
The Memory Explorer, by Laurie Mersereau --
-
Knowledge: Vast & Powerful, by Roy Roebuck --
-
"The Data Interpolator," by Paul Edstrom --
-
"HAL", The Ultimate Digital Assistant," by Carl
Keller --
-
"Personal" Power To The People!, by Damon Turnbull
--
-
The Perfect Assistant, by Robert Dickson --
Games.
A.I. (Not The Movie):
CPU-Cycle Consumers:
In Classes All Their Own:
-
The Jean-Luc School of Computer Interfaces, by
Jack Smyth --
-
Genetic Sequencing, and More, by Mike Miller --
-
Compression, by Richard Johnston --
-
Several Visions, by Bill --
-
Security, by Ralph Broom --
-
Wireless Mesh Networking, by Rob Looker --
-
"The Mark of the Beast" Redux, by Tom Williams --
-
Legal 'Enabling' of ID-related Killer Apps, by
Bruce Campbell --
-
Disaster Warning Network, by John Flanagan --
-
It's The Database,
Really, by Ed Beneville
-
'Easy-To-Use'
Common Apps Could Be The Next Killer App, by Penny
Pagliaro
I
build virtual worlds for a living. Virtual Reality
is definitely the next big thing. It feels the way
graphical interfaces were in the days of 8-bit
processors -- we could understand the potential,
but in those early days they were slow and clunky,
and took up too much memory.
Realtime VR, especially in the form of VRML
(Virtual Reality Modeling Language) has come a
long way in the last several years, but is still
in desperate need of faster processors and bigger
machines. There is great potential for VR in many
areas, but a few deserve special mention:
Multiuser VR is a brilliant communication tool.
It places very little load on networks; the end
machines do the bulk of the work, send just
position and action info, unlike a video feed
which requires massive amounts of info to be
streamed. Discrimination on the basis of race,
sex, physical disability, and physical
attractiveness become irrelevant in VR -- you
choose how you present yourself.
VR fiction is the next step beyond movies -- it
is like being able to walk around inside the
story. Sure, it is hard to do, and people are just
starting to develop the language and tools for it,
but consider how different the first movies were
from stage shows or radio. Early attempts at
movies didn't use editing, or panning. They didn't
have sound or dollies or any of the sophisticated
lighting, special effects, and other techniques
that we take for granted now. VR fiction is an
entirely new storytelling medium and I am certain
it will become a multi-billion dollar business.
We humans have filled the Earth. There are still
some unknown lands to explore, but not many,
unless you include the oceans (which we are
rapidly reducing to wet deserts). Space travel is
looking less and less viable. We can planet-hop
within our own solar system perhaps, though we do
so in dread of radiation storms from the Sun.
Interplanetary space is inhospitable; interstellar
space is even more so. Exposing our fragile bodies
to cosmic rays on long journeys is a real
non-starter. Unless we can drastically shorten
such journeys then we are trapped here. But many
of us are explorers by nature.I can already create
infinite virtual worlds on my computer. Add to
that the ability to use fractal rules to
auto-generate infinite diversity, and the growing
(excuse the pun) field of artificial life, and we
could have endless horizons available to us. We
could potentially learn much from these virtual
worlds about how the 'real' world works.
And the most controversial: Our biological
bodies are destined to wear out. They have evolved
to do that to enable the genetic diversity that
ensures adaptability and survival of life. I
really don't think we can ever 'fix' that. The
tragedy is that each of us has this wonderful mind
which has needs and abilities far surpassing the
fleeting biological capabilities that gave birth
to it. The mind is information and can be modeled
by computers; not yet, but soon.
When we are able to live on inside computers, VR
will be a natural place for us to live. Whereas
mechanical systems will always wear out,
electronics has no moving parts, and properly
designed could perhaps last hundreds, thousands,
maybe millions of years. And don't make the
mistake of thinking that living such a life would
be fake. People would still think great things,
build great works of art and marvellous tools,
love, and learn... and there is no doubt that they
would continue to be involved in the 'outside'
world. What would people be capable of being freed
from the need for food and shelter, and able to
learn for many centuries? And imagine the resource
that those millions of minds could be for those
dealing with problems in the 'outside' world.
VR
IS the next 'Killer App!'
VR,
The
'CPU Killer,' by Scott Miller --
What might the [Virtual Reality] future look like?
David Gelernter, in Mirror Worlds (http://www.amazon.com/exec/obidos/ASIN/019507906X/qid=1036396476/sr=2-2/ref=sr_2_2/104-3388988-4338319),
makes an interesting case for how this technology
might one day be used. What are the three people
in [a movie without a link] doing? (Prepare to
plunge into the imagination of Scott)...
<imagine>
What if they are three partners in the business of
creating bots and avatars for use in VR worlds?
What if what we are watching is the VR experience
as it will someday exist? What if what we are
watching is their view, their customized interface
into the software they are using to create the
creatures we see? "Mr. Light" does some sort of
high-level design, or architecting, "Mr.
Industrial" seems to implement the design, while
"Miss Magic" animates the creations. All these
steps exist today for avatar creation, but today's
GUI is NOTHING like what we see in the little
movie. Imagine if the OS of the future is an
immersive VR experience... </imagine>
Why should the Intels, AMDs, IBMs, etc. of the
world be excited?
WOW this stuff chews up processor cycles! And no
matter how powerful a machine you "throw" at the
problem, the creative artist ALWAYS needs more.
Give us a 100x improvement in processing power (a
decade of Moore's law) and it will still not be
enough. More polygons, better algorithms,
real-time procedural textures, the demand keeps
growing. And the real challenge: a better and
immersive interface. The last several issues of
your e-mail has dealt with the merging of human
and machine. "Miss Magic" acts as if she can feel
the breeze generated by the wings of that humming
bird, perhaps she can, even though the breeze
isn't "real".
Today we are stuck with limited technology that is
good for little more than playing games. What will
the future bring?
<<-Let's see if I can summarize this in a
paragraph->>
Network based Virtual Reality will be The Killer
App for driving personal (and some day, implanted)
computing and networking technology. It will drive
demand for processor cycles, memory, storage,
bandwidth, and peripherals -- this last one to the
point of pushing the human-machine interface into
implantable technology. Moore's law predicts a 100
fold increase in machine capability each decade.
Currently, VR demand is more than a decade ahead
of what the personal computer can deliver, and
will remain ahead of the curve well beyond the
middle of this century. Network based Virtual
Reality technology will create new ways for
human-human communication and new paradigms for
the human-machine interface. To quote Ted Nelson
from the first edition of his seminal 1974 book,
Computer Lib/Dream Machines; "If computers are the
universal control tool, then let's give kids
universes to control."
<<-Final thoughts->>
Ray Kurzweil has implied that it might be possible
to clone a person's mind in only five or six
decades (http://www.kurzweilai.net/).
Why would we want to do this? How could it be put
to use? Or would it only be good for finally
having an excuse to talk to yourself?
Personally, I can see living in both the real and
virtual worlds at the same time. If there were a
natural mind and a machine mind that were linked,
then one of you could take your body for a hike to
commune with nature and unwind. While the other of
you could link into the VR Sphere -- creating,
communicating, doing, even working and getting
paid for it. You could be in two places at once.
I think that the next killer app will be in the
entertainment area, namely virtual reality
environments. Coupled with lightweight, very high
resolution headsets such an application (lets call
it Virtual Adventure) would allow you to feel you
are IN a game, film or other audio visual
experience.
The next killer app, and one that will require all
the horsepower we can throw at it, is turning your
home or office PC into a fully immersive
entertainment/computing environment, ala
"Lawnmower Man." Here, the keyboard, mouse,
and monitor all go away since you will be
virtually "inside" the computer.
"We are Borg. You will be assimilated."
Of course the next killer app that will stretch
the hardware capabilities to the limit will be
"virtual reality apps," both for business and
personal use.
The first ones will be crude and simple gaming,
but as the hardware gets better there will be
incredible business applications. Imagine walking
into your companies' database in virtual reality
and having any information at your fingertips to
manipulate as you please. What about truly virtual
meetings for cyber-commuters - eliminate the need
for offices.
As for gaming, how about experiencing incredibly
violent or daring life threatening thrills
firsthand, since you cannot die in virtual
reality? What a great way to overcome fear.
[Or to suffer a heart attack or other
situationally-brought-on illness? I wonder
if there are documented cases of this, considering
the intensity experienced by many a game
player...]
Finally how incredible a teaching tool could this
be? All types of skills in sports could be
practiced over and over with no threat of injury.
The list is endless, and we are at the verge of a
monstrous quantum leap in computing.
And, not to get too much into "matrix" but how
nice would it be for the very old and the
paralyzed to be able to walk, run, and fly in
virtual. Sex will not be too far behind.
[The combination of sex] and gaming will probably
drive the advances...
Two words: 'virtual reality.' When we can use our
existing killer apps (and I think I'd add computer
games to that list) in a more intuitive, dynamic,
3D environment, I think we'll see another scramble
for computational horsepower.
The one that I've been waiting for is SVM (Short
Video Messaging). It will require lots of
bandwidth, an amazing graphics processor, and a
hungry CPU.
Generally, people don't want to send messages with
their -exact- picture and surroundings, and many
are still scared of the videophone idea since they
may be caught [in] a 'not-photo-ready' state.
Hence, the new generation of SVM will use a
predetermined video of the user during such a
'photo-ready' state, and will enhance it with the
facial expressions and lip movements during the
time of recording.
The CPU and Graphics cards to do this are
currently just above the high end of today, but
will easily be [available] by middle to late next
year. Hand-helds will be able to handle this
within 4+ years. No more worries about what
you look like.
And the bonus app that will make this REALLY
desirable, will replace the live background and
noise with a dynamically configured virtual place
and sounds. It will be particularly useful
to youths that want to fool parents/peers into
thinking they are somewhere else, since sending
the real video will always be an option.
:-)
On a related note, this technology will also allow
true virtual conferencing [over constrained
bandwidth] since only gestures will need to be
sent from each of the individuals in the
conference. Their digital signatures will be
sent during the initial synchronization, and live
updates will simply be overlaid during the
discussion. Given that large, high
resolution, flat panel MEMS screens are going to
become affordable in about 3 years, the timetable
will converge appropriately.
Fully-Immersive VR, by John M.
--
How about fully immersive VR? Picture Sandra
Bullock in "Demolition Man". Definitely a
must have. Experience concerts, rallys,
games, any sort of tech training, travel, and of
course sex, in a VR environment. Definitely
huge.
Robots? Probably another big one: the auto
that drives you to work; housecleaning/security
robots to patrol your home; personnel bodyguard
robots; etc...
Jeff, I reckon the killer app for those spare
cycles is usability. In a very few years we'll be
looking back on the computers of today, and
wondering how we ever managed to use them. Think
of Scotty back on Earth (I can't remember which
movie it was) when he prepares to remonstrate with
a Mac because it won't obey his (voice) commands.
Sure, he then uses the keyboard, and gets the Mac
to do things you and I couldn't, but that's the
fiction bit of the SF movie.
It's too easy for techies not to recognise the
issue (I'm sure it's an issue you do recognise).
I'm no slouch with PCs, and at work I'm one of the
guys the others ask if they have a problem. I like
to get into the guts of things, and read the
hardcore journals. But now, even with XP, when I
check the processes in Task Manager to find that
servicehost has multiple instances, and I can't
figure out which is causing my problem... then I
bang my head against the wall.
PCs still can't tell you what's wrong with them,
or how to use them. The reams of 'hot tips' in
magazines, and semi-expert newsletters attest to
the fact that they are usually not set up to get
the most out of them, or sufficiently personalised
for most people. But most people can't understand
the instructions on how improve things.
So if someone can build these things into an OS,
or if someone can produce an app that will sort it
out - they'll be in the money.
The killer app in 2010 may involve holography.
Imagine a "PDA" being able to "project" the
life-size image of Brittany Spears or Justin
Timberlake dancing to one of their songs in your
teenager's bedroom such that he/she could dance
along.
Or it could project an image of a mechanic
repairing a complicated airplane problem such that
the learning mechanic could follow along
step-by-step.
Or it could project a skier doing some turns while
you follow him on an interconnected device that
tracks your movements and makes suggestions for
"up-weighting earlier".
If the networks of the time can provide the
bandwidth, new holograms could be accessed via the
net and commerce could support the "killer app".
My idea came from:
Napster
Keyboard projection technology
Teenager buying enthusiasm
MTV/MP3 popularity
Personal life experiences
I think the next killer app that will drive the
need for computer speed has already been invented
[(and seen in the form of special effects] in
movies - take the original Star Wars as one
example); it's really just waiting for the display
device - and perhaps the horsepower - to make it
practical.
For lack of a better name, call it "Holo-CAD
Drawing," where you can get a true 3-D view of
your designed object, along with the ability to
zoom in and get a detailed look (using routines
that adjust the image so it still looks smoothly
real.) It will be the next boon for
Engineering.
Imagine creating a product and being able to
display it (literally) exactly as it will look
when built. The obvious game implications are also
there. The real limit to this, currently, is a
display device that would permit a seamless
display from any angle, even top-down. Figure that
out, and you have something that will really drive
increases in processing power, once again.
I
bet your average size holodeck or even a full
length feature 3-D holo-flick action movie could
easily chew through at least 11 or 12 gigahertz of
processing power and 8 - 10 terabytes of memory.
[Vast amounts more of processing and other
power, from what I've been reading. JRH]
Or how about a transporter-like "3-D FAX machine".
That could use a good chunk of power as well.
[Remember "stereolithography"? Such 3D
printers already exist (see
http://www.compaq.com/rcfoc/19991018.html#_Toc464536634),
so extending them "online" would seem feasible.
JRH]
Just look to your local independent TV station
running on series or another of Star Trek and I
bet you can find at least 10 items that could chew
up that much processing power, and more.
No matter how much processing power we have, we
will always find ways to use it. Pushing the
boundaries is what makes us human.
With advances in broadband that should take place
during this time [period], I ... envision a few
ideas:
"Shop-at-home", a virtual-reality approach to
shopping which provides a very high resolution
experience for shopping. The recently
announced ability to touch over the internet would
make fabric choices easy. New algorithms
which provide mapping a person's physical
characteristics would provide the consumer with
visual feedback, while providing the alteration
shop specifics regarding the customer's specific
needs.
"Home movies", a video mapping and animation
application to allow people to take home movies,
perform visual effect editing, voice-over and
Foley dubbing, and create stories in a very short
time. This would prove especially valuable
for gifted and talented amateurs (producing at
least early "Star Wars" quality special effects as
standard). Create your own exotic vacation
experience.
"Design your own home", a virtual-reality approach
which provides true real-time walk-thru with the
ability to move walls and furniture around with a
simple "push" or "pull". The walk-thru would
use VR eyeglasses, providing the average consumer
with the abilities that architects see today as
cutting-edge. This could also prove a boon
to the professional designers and homebuilders.
"Virtual Museum", a virtual-reality approach to
viewing antiquities. Museums would have
their art scanned in full 3-D. Customers
would pay a fee to enter the museum, and the
visual effects would give them images that rivaled
life. This would also provide the ability to
linger at any exhibit for any length of time, and
to view the collections that are not usually on
display. Could easily be extended to the
scientific community for collaborative research
(again, using the emerging "touch" technology".
Naturally, "virtual porn"... With the
virtual-reality visual effects, combined with what
can only be imagined in the "touch" arena...
Not sure how far you wanted to envision the
possibilities! <g>
I can imagine a few more, but it seems the future
"killer" apps will focus more around consumer
experience and entertainment than around utility.
Witness the browser in the refrigerator (http://reviews.zdnet.co.uk/review/28/1/1600.html).
John provides his answers to the next Killer Apps
by quoting from the Internet2 faq
(http://apps.internet2.edu/html/faq.html),
written by Ted Hanss, Internet2 Director for
Applications Development:
What's the "killer app" for Internet2?
The fun thing about participating in Internet2
is working with all the creative minds who are
working to create the next generation of
applications. But, there's no way to tell what the
"killer app," if any, might be. Instead, we like
to describe four killer attributes that you'll
find in the most compelling applications. The
first is interactive collaboration environments,
where you can truly interact with others without
the barriers of distance. The second is to provide
common access to remote resources, such as
telescopes and microscopes. The third is using the
network as a "backplane" to build network-wide
computation and data services, such as those under
development in the
Grid. The fourth attribute is displaying
information through virtual reality
environments—moving from statics graphics and
images to moving, three-dimensional animations.
No, really, what's the "killer app"?
Well, the area that will provide the widest
benefit and largest aggregate use of the Internet2
network capacity is digital video. Video-based
applications cover everything from video
conferencing to on-demand content to remote
control of microscopes and other instruments.
This one's been on my mind for years. Given
more-or-less ubiquitous broadband capability, I
think people at work and at home will discover how
much communication is enhanced by others' facial
expressions.
This will lead, in turn, to major changes in
business and educational travel decisions, and I
seriously believe it will change our culture to an
extent not seen since the telegraph. Where will
processing power come into it? Echo
suppression, which is the current bane of PC-based
videoconferencing.
Once broadband is more pervasive, even more so
with the upcoming Internet 2, I think VIM (Video
Instant Messaging) will become a killer app.
With digital USB stereo headsets, and USB2 &
FireWire web cams, we will have crystal clear,
full motion video and audio. This will surely
replace the home phone (but not cell phones, of
course). This could be for both one-on-one or
conference calls. This is especially useful for
friends who now live long distances from each
other, yet want to stay connected. This could be
integrated into one's computer or used as a stand
alone video phone device. All of this with no long
distance phone charges. And in time, this could
evolve into your MMORPG where you mentioned
lifelike virtual 3D coffee shops. People would
have no excuse for not keeping in contact. And
this would be a ... lot cheaper than an airline
ticket.
On the other hand, I really like the TiVo, Replay,
Digeo/Moxi devices. But I would rather incorporate
this into the computer where I can easily upgrade
(add) a larger hard drive or TV tuner card. This
will be the entertainment server idea like Windows
new Media Center, but without any imposed
limitation. Ideally this would have two HDTV tuner
cards (one to watch and one to record, or perhaps
record with both at the same time), 2 to 4 120+ GB
hard drives, a DVD burner, wireless Ethernet (a, b
and g) for sharing amongst other devices
(computers, TV's, PlayStation 2's, stereo's, etc.)
and Bluetooth for connecting wireless keyboards,
tablets, etc. Couple this with a one click
TitanTV.com-like scheduler and you have a simple
DVR/PVR.
(Currently, some of these capabilities are already
in bleeding-edge products, although not all of
them in any one product, and none of which have
yet been leapt-on en-mass by consumers. But
this may be a leg-up for Mark's ideas.
"Digital Buddy" is a life size holographic multi
user, real-time, visiting software and interface.
It allows you be with friends and family from far
away, yet with the visual and auditory feeling of
reality. Since it is multi user enabled, you can
have several friends "over" at the same time to
watch movies and play games. Your friends appear
either in your monitor glasses, contacts, or
projected into the room where your holographic
projector is placed. Each user has the ability to
provide a width / depth / height of their
projection and reception to include the necessary
surroundings.
Forget your phone; be with your loved ones no
matter where they are on the planet.
Virtual Tours, and More, by Byron Law --
I imagine that one killer application that is
going to drive processor speeds and storage
capacities is going to be immersive video.
That is, video that is recorded omni-directionally
and subsequently is viewed omni-directionally.
There are many that are working on and perfecting
the recording side of this, but viewing is still
relegated to a mono-directional monitor with the
viewing angle adjusted by mouse or keyboard,
rather than a more natural user worn system that
detects head or eye movement to adjust viewing
angle within the video. Once immersive video
is perfected to the point that the user is under
the illusion (at least visually and aurally) that
they are actually at the remote site, I think many
will start subscribing to various tours and
experiences.
How often have you heard the complaint of those
who visited the Smithsonian that they wish they
had a couple of weeks to spend in there?
With immersive video, they will have all the time
they want (within their subscription) to "roam"
the Smithsonian (or any other museum, tour, or
countless other experiences exploring every corner
of the globe and beyond) and from the comfort of
their own home. Field trips for schools will
be as simple as scheduling a short walk down to
their library and donning their immersive video
headsets. Schools will have a much greater
variety of experiences to expose the children to
(that coordinate with a greater variety of
topics), with reduced liability and safety
concerns. Maybe someday we might even have
sophisticated immersive outer space based
telescopes (with the requisite blocks looking back
towards Earth against peeping Toms). We
could use more astronomers scanning space in every
direction possible. We might even see law
enforcement using the footage from the blocked
portion to better solve crimes.
Here are a few sites I know of that deal in
immersive video recording technologies:
http://www.fpvideo.com/
http://www.behere.com/
http://immersivemedia.com/menu.html
http://www.ipix.com/
http://www.spincam.com/
http://www.iqeye.com/mkt/immvid.htm
The
In(puts) and the
Out(puts), by Bart Wessel --
What we are going to use all the processing power
for? Human interfacing. (Aren't we
already?).
Let's face it: most of us deliver input using
cumbersome mouse and keyboard manipulations (many
of us still cannot type fluently, and plenty of us
get to suffer from RSI). With respect to output:
we get to sit all day long in a fixed position
relative to a display device that is made up
out of individually discernible dots. Ha!
My guess is that we're going to see input
recognition (speech, gesture, facial expression,
...) and output generation (projection,
3d-holographical vision, ...).
If only I could put my facial expression at this
very moment into words, you'd see what I mean.
Imagine, say, by the year 2012, with even
pre-molecular processors, the p-o-r-n-o and
computer gaming industry of 2002 will seem like a
blip on the screen when one can walk into your
local Fry's and purchase the latest 'Whoever does
Dallas' "tactile", (T-DVD?) and instead of a joy
stick, well you get the picture.
In addition to just talking through the PC's
[microphone and speakers, for Internet-based phone
calls], let the PC be a 'smart' answering machine
- dealing with taking messages when I'm not
available and managing them better than just
'listen and push the delete button' on my dumb
answering machine.
Maybe customized messages for when specific people
'call'. And of course conference calling, stored
phone lists integrated with public directory
lookup, etc. Not to mention moving into video in
addition to the audio.
I don't have a catchy name, but the next killer
app is going to be the use of full-frame,
full-speed video in all the other killer apps!
This coincides with the processing power to handle
such media, as well as the increase in web
bandwidth to handle it.
As for the killer app, your newsletter suggests a
version in the HAN piece
(http://www.theharrowgroup.com/articles/
20021104/20021104.htm#_Toc23925814).
Coupled with advances in storage, we could create
the personal recorder, a Tivo for each of our
lives. With an eyeglass camera and audio recorder,
we could easily have exact replications of the
vital times (from many perspectives). Sure, it's
fun at Mardi Gras but we could also precisely
record a business meeting, political event,
wedding, first date, what have you, and the
processing power would enhance not only the
recording but also the search and recall. You
never forget anything. What would this do to
marital quarrels?
If we could ever get past the bandwidth/regulatory
issues, we could open a much larger can of
possibilities but that is a legal issue with no
end in sight. In the short term, though, there are
strong possibilities for adapting mesh network
communications in unrestricted bandwidth ranges to
at least entertainment applications, such as
interactive games in a crowd (or even a mall).
Couple this with the processing power and you
could have some potent fun. Ideally, you could
also create a wireless mesh Internet but that
needs large-scale population density.
(This
suggestion was short and sweet ("life recorders"),
but Swift is talking about sufficient audio, video
(and perhaps more, like GPS) inputs to a massive
data store, that will duitifully and invisibly
record every aspect of every moment of your life.
(I just hope it comes with a "Pause" button...))
The Killer App, nick-named simply "Real Life", is
real-time video rendering--without a trace of
blockiness or delay.
This technology would transform any device with a
screen (laptop, tablet, palm pilot, cell phone,
digital cameras and recorders, even eye glasses)
into a two-way video camera that one could hold
up, pan around and instantly transmit the video to
another user. The ultimate extension of this
technology would be for one person to wear eye
glasses on a vacation and be able to look around
and transmit every last thing he or she sees,
continuously, to a recording device back home or
live to another user wearing matching
glasses--such that the second person forgets
he/she isn't seeing these sights first hand.
This would completely transform the movie, home
movie, news, and TV industries, as well as
military reconnaissance and battlefield tactics.
There would be many linkages to Internet,
advertising, on-line gaming, on-demand downloading
of TV shows and movies, etc. More simply,
one could even hang "pictures" in one's home or
office that are linked to on-line cameras at
national parks or art museums to provide a
"window" on art, or on the rest of the world.
Home and office art that changes with the viewer's
whim. Suddenly, any device at all could
double as a TV screen or a videophone.
Such rendering would consume tons of bandwidth and
memory, forcing the development of better
graphics-rendering cards and much faster CPUs/chip
sets, bigger and faster memories, and wireless
bandwidth. But live video rendering would
have a market that subsumes every last device with
a screen. A couple of pins on the frame of the eye
glasses could allow the wearer to stop, start,
freeze-frame, black-out, or add effects to
whatever he or she was seeing/recording. One
could even "import" more interesting, exotic
backgrounds to make the average walk or ride to
work more interesting. Truly seeing the world
through "rose colored glasses."
Other advances might be HDTV format for these
glasses or video in general; some graphics
algorithms that corrects for haze, or overly
bright sunlight, or dust/pollution; and built-in
binocular function for distance seeing, a la Luke
Skywalker in the first Star Wars.
Another, much more simple technological
improvement would be a computer with "instant on."
No more waiting 2-5 minutes for a computer to
boot. This would make such a great selling
point--think how many minutes are wasted each and
every day waiting for your system just to boot up.
(Simplest thing here would be a computer that you
never had to turn off, that went into some natural
safe mode or hibernation state that kept it secure
from the rest of the on-line world).
Let's call it the "Personal Companion" ( PC ?)
A voice recognition, interactive, multi-purpose
virtual companion. A program that could have
a custom or randomly generated personality that
would learn a user's preferences likes and
dislikes. After a "get to know you" period the
"PC", in addition to mundane tasks such as
schedule planning, important date notices, Web
ferreting, house monitoring etc., would also sense
any tension in the user's voice and prompt him/her
into, for example, conversation designed to
relieve stress.
The PC could also periodically run a series of
health questions [on a schedule] depending on the
age/health of the user.
And the PC could design custom entertainment
packages. For instance, it could automatically
download preferred music, movies, Web site
information, TV shows and sporting events, which
could then be enjoyed at leisure. Anyway I think
you get the idea.
I think the next killer app will be "true voice
response". By this I don't mean the poor voice
recognition we have now but true voice response
where you can dictate a letter, stop in the
middle, order a switch to your browser, look up a
word or phrase online and then switch back to your
letter - maybe with copy and paste into your
letter - and go on dictating and have the letter
be done in perfect format without you having to
dictate all the punctuation -- as people who have
their secretaries type from a recording do now, or
their having to touch a keyboard or mouse.
I think this will take all the CPU power we can
throw at it, and have the advantage of allowing
quick easy entry of data into machines that for
one reason or another don't have keyboards or
mice. Think of adding a memo or appointment to
your Palm device by whispering to it. Also, with
our aging population, fingers are becoming less
agile, so there may be a growing market there.
The
Butler, by John Smith --
This is probably not the NEXT killer app…but I
think it's on the way.
A major step towards improved ease of use could be
the trigger to get people motivated to buy some
new technology. Integrated into the various
household appliances, communications devices and
A/V technology with truly reliable voice
recognition and a (seemingly) intelligent response
this would be something that would initially
interest and amuse users and eventually would lead
to this becoming the only way to interact with the
technology around them.
And so, I would like to introduce you [to a future
ad for] -- "The Butler:"
"Have
questions answered from anywhere in your house.
Find useful information without sitting at a
keyboard. Get things done while you're busy
elsewhere.
How?? Get "The Butler" to do it!!
"The Butler" is the new software for your home
that responds to your voice. Ask him questions and
he'll find the answer and tell you. Give "The
Butler" orders over the phone and he'll carry them
out for you. "The Butler" can talk to you and keep
you up to date with what's happening at home. Have
your mundane tasks handled by someone who always
ready to do your bidding. Buy your new computer
and "The Butler" today."
The
"Other" Butler, by Max Rible --
One obvious 'killer app' (which will require
advances in [other] states-of-the-art to
[complement increasing] processing power, but
which still might be bottlenecked on
processing power and memory) would be really
effective voice recognition and language parsing.
It'd be a real mess in a "sea of cubes"
environment (can you imagine a room full of people
all trying to talk to their computers at once?),
but it'd be superb for the novice home user who
wants to just phrase a verbal query rather than go
through all the tedious steps of clicking through
a user interface. (Consider the speed of speaking
"please open the draft Harrow Report for next
week" compared to File/Open/navigate
directories/...)
If AskJeeves.com weren't already taken, I'd
suggest naming this app Jeeves, but I guess I'll
have to just use the name of Batman's butler,
Alfred, instead. Naming your computer would
be an effective way to let it know to pay
attention to your next sentence; you don't want it
to trigger off of "Sure, I'm planning to format
the hard drive next week", but [even] "Alfred,
please format the hard drive" would be worth
prompting about.
[Note that there are already -fair- voice command
processors available, and some of these early
applications do allow for 'naming' the computer as
a "verbal key." But this generation is still
far from 'good enough' unless you are in a
consistently quiet environment and you remember
the explicit 'command phrases'... Yet I
agree that all of these issues will likely be
surmounted.]
The
Butler, Times-3, by Kim Mains --
I've become most
interested about the possibility of the different
Robot technologies from Sony/Honda (ASIMO)
meeting up with some of the
work done at MIT, especially the
Kismet work, along with the introduction of
Bluetooth, home automation, and home "always
on" cable connections. Of course speech
recognition is thrown in there somewhere.
How about asking the robot what the weather is
going to be like and getting a response pulled off
of Weather.com. Technically it's not that
hard if the robot can understand the request,
knows where to find weather information on the
internet and can do text to speech. The
robot with wireless internet access via WiFi or
Bluetooth could easily do this without any
apparent effort.
Or [how about using robots for] operating the
appliances and other devices around the house?
The robot could be the universal control panel to
everything. You could ask it to turn the
heat up a bit (it could transmit the request via
X-10 RF if it had to). It could change the
channel on the TV; turn on the hi-fi to your
favorite music; and all without lifting a finger
via wireless commands. IF, [that is,] you
have wireless support in your appliances...
It could also read your email to you amd, when the
speech recognition gets good enough, take
dictation.
Could [this herald]
the return of that 19th century institution... the
butler?
One curiosity that I
have already encountered when talking with my
friends: I mentioned the Electrolux robotic
vacuum which has the perception altering ability
to plug itself in when it's batteries are getting
low. I have someone clean my house so I got
the inevitable suggestion that I use the robot
instead. Then, the person rejected the
idea... after all, the human is cheaper [today, at
least...]
One killer app which might help drive CPU power
forward is the old standard of voice recognition
software and perhaps even having computers talk to
you. This could be like Star Trek’s
computers where you talk to the computer and it
responds to you by speaking to you in normal human
voice unlike today where it is mostly a mechanical
sounding voice. There are many opportunities
with voice/speech options. I’d like to say
“Computer On” and have it boot up, then have it
open the e-mail program and read the subject lines
or the news headlines of news e-mails, etc. etc.
The [Killer App] that I'd spend the bucks
on is a true voice recognition interface.
Call it "SpeakEasy" - it would free me from having
to use the keyboard so darned much, and more
importantly, allow a smooth flow of words to be
processed by my PC.
We've been hearing about VRI for years, but it has
always been sketchy and costly, and way off the
mark user-friendly-wise [too much set up time, and
too many errors]. SpeakEasy will allow me
and mine to sit down for a few minutes, give the
PC a 2 or 3 minute sampling of our normal speaking
voices, and SHAZAM! provide us with an easy
hands-free interface.
What a boon for writers, secretaries, and anyone
else who needs to process their thoughts. I
know that when I am in a stream-of-consciousness
mode, I cannot type fast enough, and usually loose
something before my fingers can get it all down.
Further, if it was set up to process voice
commands, well then that would just be the
icing on the cake! Then the blind and
physically challenged communities could really
take advantage of the power of the internet
and PC, without the cumbersome interfaces that
currently exist.
I think the next killer application that we are
all hanging out for (and which will undoubtedly
demand lots of processing power) is computers we
can talk to, and interact with to carry out normal
daily tasks.
This is of course again more Science Fiction, but
it feeds our innate desire to have a slave waiting
to carry out our every whim with only a word or
two. I'm not just talking about dictating
memos, but the interactive computer that controls
our car, our home, and our entertainment system,
and can also teach us about the increasingly
complex technology that we simply never have time
to fully learn.
This more than anything else will also help bring
advanced technology out of Nerdsville and open up
advanced computing to everyone who can communicate
verbally.
Superb, And
"Learning," Speech Recognition, by Jack Lipscomb
--
I think that we are working our way toward real
artificial intelligence. The new Killer
Application will be SUPERB SPEECH RECOGNITION.
It would be used in word processing, computer
control, and general interacting with the computer
including Internet searches. Later, this will be
married to the ultimate killer application
"Learning Artificial Intelligence."
These would be programs that pay attention to
their environment and learn as we do. They would
be able to interact with each other on the
internet and learn from each other. They
would listen to us and talk with us both in person
and on the telephone. They will learn to
perform complex tasks such as "Computer, please
research the state of the art in lightning
protection systems. Check with me on the cell
phone when you think you have a handle on it."
They would become gainfully employed.
But first we must develop the next Killer
Application --SUPERB SPEECH RECOGNITION .
Making Computers "Disappear,"
by Jeff A.K. --
I think the killer apps need to assist in removing
the current I/O devices:
Voice recognition and translation which really
works!
Also voice capabilities which can filter through
pauses, strung-together words, slurring, dialects,
etc., without error.
Or language translation in a many-to-many format
without error, or with very few.
Voice-capable GUI navigation or the creation of a
GUI which would not require a pointing device, but
will be more intuitive to voice commands.
Chip-based security for both storage and
transmission.
Anything which makes the computer "disappear" will
surely help.
Actually, I don't think the "next killer app" will
be just one app...
I think it's going to be voice recognition, which
of course will be applicable to MANY apps.
For it to be useful, though, I think it has to
have nearly a 100% recognition rate, and we know
how tough that is. However, with enough
processing power...
How about REALLY effective voice recognition?
Something that can handle naturally spoken
sentences without having to pause between words.
THAT would require some pretty quick number
crunching.
I propose the following four ideas:
One of the biggest restrictions in computing in
our mobile and workday-shifted workplace is the
physical dimension and interface of our computers.
With various computer output replacement options
such as goggles, monocles or heads-up displays,
there is only one currently practical hands-free
input option, and that is voice.
While that technology is coming along nicely, it
does not allow for privacy. I believe that a
much needed killer app that will likely consume a
lot of processing horsepower will be the ability
to decipher brain wave activity as the input
medium. Work is being done in various areas,
but it is still a way off. Once that
technology is developed, however; the whole world
of communications, travel and entertainment will
change dramatically.
The computer entertainment industry is based in
part on making more realistic computer graphics.
With more brute processing capability, holographic
imaging may become the next evolutionary
development. It could be used in many areas,
but the entertainment industry, including gaming,
movies and adult entertainment, will drive the
funding for this technology.
[Indeed, this promises to be a real CPU-killer.
From the Nov., 2002 MIT Technology Review:
"Still the diffraction pattern from just one
high-resolution hologram can easily use up more
than a terabyte of data—enough to fill 1,600
compact discs. A moderately flicker-free
holographic video would require at least 20 such
holograms per second. Clearly, churning through 20
terabytes worth of information every second would
require extraterrestrial technology: today’s
fastest PCs operate at one- hundred-thousandth
that rate."
Details at
http://www.technologyreview.com/articles/freedman1102.asp
.]
While the "Star Trek" "Holodeck" is a pretty cool
idea, there are a lot of technical difficulties
associated with trying to develop the matter
replication needed for tactile feedback. But
what if the ability to capture the brain waves for
input to a computer could be reversed and the
output of a computer could trigger inputs directly
into the brain, bypassing the requirement for
goggles and VR tactile gloves or suits? This
would definitely change the entertainment and
tourism world as we know it. The processing
horsepower would be immense, but wouldn't that in
and of itself drive the computer industry even
harder?
There are a whole new series of plush toys being
advertised for this holiday season that are filled
with animatronics and priced a lot lower than the
mother of all "toys"- Sony's "Aibo". With
greater acceptance of toys that interact, the
drive will be on for low cost and ever more
responsive toys. The "A.I" movie Teddy Bear
is not that far off if you think about what could
be done by combining an "Aibo" and "Furby" into a
plush responsive Teddy Bear. I see a
child-friendly, human-machine interface as the
impetus to drive future simplified interfaces.
And as these children grow up, they will expect
that same level of simplicity and sophistication
in their home entertainment and workplace.
The Invisible User Interface (IUI) - i.e.
speech/video recognition.
Vertical Intelligence - i.e. a series of
intelligence modules that outperform must humans
at a given subject. Full (horizontal) human
level intelligence will have to wait a further 5
or 10 years, but hey, that's just a blink of an
eye...
According to Sony, NEC, Hitachi, Honda, Husqvarna,
iRobot, ActivMedia, Evolution Robotics,
GeckoSystems, the United Nations, DARPA, et al,
the next BIG THING is "personal robots."
Maybe it is just because I am now messing about
with Video, but much better Video editors and DVD
burning software would be a "killer" app since
people have now shifted to video from 8MM film and
can do "digital" things with it. We all could use
a 10 GHz processor with a Terabyte of disk.
If you haven't, you will.
The
More Things Change, The
More They Stay The Same, by Chris Denver --
Yesterday:
The first PC I had on my desk was an HP150
touchscreen - followed not long after that with an
upgrade to an HP Vectra (6mhz - 4meg ram -
20mb HDD - A veritable monster !) Back then, the
killer application was MS Word for DOS 2. It
provided pretty good response - used to take me
about 3 seconds to open a word processing
document.
Today:
These days I have a 1.5ghz P4 - 384meg ram - 40gig
HDD The killer application today is MS Word
10. It still provides pretty good response
- takes me about 3 seconds to open a
document !
Tomorrow:
Although I know better, I'll probably go set a 15
ghz box, with one terabyte of ram and a limitless
universal internet-based virtual storage
system. The killer application will still be MS
Word 25. It will still provide pretty good
response - it will open a document in about 3
seconds!
BUT HANG ON A MINUTE !! - Don't
say "ahh.... MS Word 25 will the version
with reliable voice-to-text processing." An
industry expert once told me that voice-to-text
was the next big technology - and that was in
1977!! Since then it has been said every two
years. Without fail!
For my money (and for a LOT of Microsoft's R&D
money), the next "killer app" might end up being
an INTERFACE more than an application. Voice.
I can easily think of hundreds of applications of
having a true voice interface with my computing
devices. Sure - there are times when voice will
NOT be appropriate (for example, when I am having
a private chat with my lover during a boring
Monday morning sales meeting). However, in the
vast majority of instances I can see myself
preferring voice to typing or handwriting. A few
examples:
How I interface with my car while I am driving
("play all Lou Reed tracks on my hard drive", "set
cruise control for 100", "lower the roof", "call
my boss at his unlisted home number", "plot the
best route home based on current traffic
conditions", "read my new email"…etc)
How I interface with every electronic device in my
home ("set climate control at 21 degrees", "lock
all doors and windows", "record all television
shows with the word "wacky" in the title", play
all Lou reed tracks..", etc)
How I interface with my PDA under normal
circumstances ("bring up all files from last
calendar year with the word "hailstorm" in the
title", "show all journal entries for the last 12
days", etc)
You get the general idea. The main problem with
delivering a solid voice interface on all
computing devices is one of processing power. We
need more crunch. Rock on, Moore's Law!
It's simple. The next killer app will be some form
of "The Big Book." It will be a portable digital
copy of every piece of information known to man.
Imagine a PDA like device that could contain every
book, every newspaper, every movie, every
photograph, every web page, etc. With digital
storage density increasing by leaps and bounds,
some day it will be practical for everyone to own
their own copy of "The Big Book." At that point,
it would be tedious to purchase information in
parcels, like buying a book, or renting a movie.
The business model will be more like "pay per
view". You will own a copy of all of that
information, but you will pay fees to access it,
just like we pay ISP fees now, to access the
internet.
Home Entertainment Center, by Joel Millett --
I would envision the next Killer App to be
software to enable the concept of "PC as
entertainment center." In other words, the
PC will control the household's music, movies, and
TV recording functions now handled by separate
components (stereos, VCRs/DVDs, and hard-drive
based recorders) and distribute them throughout
the house. One barrier to this vision might
be the lack of appropriate cabling in most houses.
Existing coax cables may be part of the solution,
but hardware makers may have to get innovative
with the use of wireless transmission to make this
vision a reality. As far as a name goes, who
knows? We can only hope the hardware and
software companies decide on a standard format
before trying to implement this concept!
Your question re killer apps
left me with an "I know there will be something,
but I don't know what it is," feeling. I can't
think what all pervasive function (i.e. relatively
simple to be attractive to users) would require
enough processing power to justify the upgrades.
However, as an interim, it may just be video
processing of home movies.
[A longish introduction follows to illustrate why
this process is so complex and demanding of
resources...]
When my father was alive he took quite a bit of
8mm Cine film which, over time, was spliced
laboriously into 5 reels running about 25 minutes
each for projection as 'family entertainment'. He
also recorded a taped sound track to sync with it.
Needless to say it all took a while to set up and
didn't get out of the cupboard often, especially
once the novelty wore off. About A year ago I had
the film converted to video. However the
soundtrack was on reel to reel tape and the tape
player I had was unusable because all the rubber
drive bands had turned to goo over the years. It
happens with that model and you can't get spares.
So the converters tried to do the audio tapes for
me but the sound quality was poor and the ones
they tried didn't seem to sync. with the film.
Anyway, to cut to the chase, I discovered that I
could more or less sync the video and the audio,
but there was a long lead in on the audio. Further
the commentary they tried was an early version (I
did not realize that my father had produced a
later update) which explained the apparent poor
quality AND the mismatches - the film had broken a
few times over the years and bits were missing! I
found a friend with an old reel to reel tape just
about working, downloaded the audio and realized
that, with some playing around due to
stretch/shrinkage, I could get the whole thing
into a presentable state which could be reprinted
to video AND be retained as a PC/DVD based
entertainment. Great! - go immediately to the
interesting bits.
No need for it ever to be 'in the cupboard again'.
As part of this plan I bought a new PC - 2Gh P4,
512Mb Ram, 120Gb disk, DVD/CD-RW, etc. OK, the
2.2Gh had just been released but was not worth the
extra price premium really. (This was last April!)
It came bundled with the LE version of what of the
top video editing programs .... The Video (2hr 20
mins) to AVI download went OK, hardly any dropped
frames, and produced a 30Gb file!. No backup for
that one then. The Audio Tape download resulted in
a 1.2Gb file.
Just this last few days I managed to make time to
revisit the project with the skills I had been
picking up over the months. I finalized the
Video/Audio sync. which was easy enough but time
consuming, being effectively and analogue process.
So now to write the output files. Hmm. A few tests
show that I have a few problems to resolve yet - a
bit of a learning curve as to why picture starts
as the right hand side on the left of the screen
and slowly moves right and then disappears off the
right hand side over a period of about 8 minutes.
There's something I don't know - and the user docs
are no help it seems.
[Focus on the END-user.]
Now, I may not have got the optimum compression
codecs applied and so on but the output file runs
to about 36Mb for 9 minutes and took about an hour
to process the preview file. Conclusion? Home
movie production, as an extension of family
photo's, could be a very attractive and relatively
affordable hobby for many people. Whilst the kit
to create is fairly expensive, the 'razor blades'
are relatively few. No film stock and processing
costs as such, for example.
But currently the software to manipulate the files
does not seem to be that end user friendly. I
reckon dedicated hobbyists or semi-professionals
will be happy-ish, but not Joe or Jo down the
street. Still, now that [ever-more] "PC Power is
around, it would be worth developing the software
user interface. Joe (or Jo) could spend a little
time editing then, [then join] the family off to
bed giving a few hours of system 'free time', they
could set the 'print' process running and expect
it to be finished by the morning.
Now this app. may not be such a new idea, but I
think it is possible that it will get a new life
and broader user base once it becomes realistic to
use. It also sort of fits with the development of
video phones and so on.
As the mobile phone outfits have a lot of money
invested in technology looking for new problems,
they are somewhat focused on this idea and are
marketing quite heavily - so public awareness is
being sparked. Interestingly my new PC, chosen for
it's spec/price combination as much as anything,
is a Sony Viao desktop - hence the focus on Audio
(though not too much I felt, lots of software
working around copy control but no separate
specialty soundcard) and video/digital pictures.
Maybe the app to become all pervasive will be one
pushed by the consumer electronics branch of the
industry rather than business applications.
Potentially it's a much bigger market for power
devices.
[And Then There's GPS.]
What can you say about the trade-off between
personal convenience/safety and loss of liberty?
You mentioned the UK CCTV cameras. We also have a
rash of traffic 'safety' cameras (for safety in
most cases read 'speed') many of which are forward
facing (front license plates required on vehicles
in Europe) and so they take a picture of the
driver and any passenger, as well as the vehicle
and plate.
Mobile Laser devices with video attached have been
doing this for some time, but now fixed cameras
with infrared 'flash' are becoming very popular.
(Not to drivers, though, who are seeing tickets
increasing by about 10 fold each year!). Other
commonly used [traffic] installations are the so
called SPECS cameras; they 'pattern match license
plates video'd at 2 locations, calculate the
average speed of the vehicle based on the elapsed
time (automatically, it is claimed), [and then]
send a ticket. The video, once again, shows driver
and passenger, and is of course stored as evidence
in case of dispute.
Add in location tracking via mobile phone signals
and build the links between the various databases
and .... well, your guess is as informed as mine,
probably more so. GPS devices which transmit
location, no matter how beneficial that might be,
are another part of the overall capability of
various agencies to undermine what we think of as
our 'personal liberty'. I wonder if such a thing
really exists or ever has existed? Either way
humanity (as a whole I suspect) will either have
to collectively accept or reject the concepts. I
can see too many negatives to a mixed approach.
As for the embedded chips concept, used well it
sounds great, used badly it would be horrendous.
You would never be mugged (knowingly) for your ATM
card, but I guess you would be dragged to the
machine...
As for all the software working together - nah, it
will never happen in my lifetime or that of my
kids, I reckon. Not at a level that would feel
comfortable anyway. Imagine that you live in a
world where peace and love reigns as everyone
knows everything about everyone else. No personal
liberty? Or total personal liberty since all
anti-social events could be controlled and
eliminated? Utopia. However, I would be
worried to be in the vanguard of such
developments. To be in a minority of the world's
populace that had become totally dependent on
technology for its support systems would not be my
idea of heaven. There would be nothing to prevent
invasion by the masses of unchipped world
citizens! So the first thing to do would be to
make a MASSIVE investment chipping (and therefore
gaining control over) everyone else ...
Hmm, methods may change with technology but human
nature probably does not. Mind you, the controls
would require an awful lot of computer power!
Could be the next Killer App. (literally). Sounds
sort of familiar again - Aldous Huxley, 'Brave New
World,' but with added technology.
As our society ages you will see a killer app with
medical sciences.
Imagine an appliance hooked into your PC that when
you breathe into it, or stick your finger into it,
it will give a comprehensive diagnosis of any
conditions you might have, blood pressure, sugar
and Ph levels, etc. Then list all your treatment
options and maybe even make the doctor aware of
your status.
Video Recognition, or 3 dimensional object
recognition ( 3DOR ) and its cousin 3DMR (3
dimensional motion recognition):
Following character recognition (finally fairly
good with the latest ones out there), handwriting
recognition (getting there), and voice recognition
(still growing - perhaps at 3 to 5 Ghz.), 3DOR
will allow a vast number of automation projects to
take off. The world of robotics will finally
become a practical reality, and we will be able to
properly take care of our elderly population.
Not to mention fantastic video game controls.
I think the killer application will actually be
built into the operating system and/or file
system.
Today it is just too easy to get data, but also
too easy to lose it. The "Memory Explorer" feature
will not only index every word in a file but every
feature used and every change you made. Actual
file names will become optional because you will
no longer need to remember the name of the object
- just a few faint recollections about it will be
enough. The "Memory Explorer" will also spend all
those spare cycles (while you are doing mundane
things like word processing, games, or Internet)
looking for correlations between all the other
objects so that if you even remember a related
file it will be able to prompt you with a
suggestion. Furthermore, the "Memory Explorer"
will constantly be scanning the web looking for
correlations that you might not have even thought
of that could be used to tie your information
together. I expect this will run in the background
at all times and thus consume more cycles than any
other application.
This might seem like it's more of a business
application, but imagine if a typical person,
anywhere on the planet (or off it), could know
everything they need to know from a local to
global perspective, when they need to know it, and
had a complete record of their past situations as
a learning aide. This is a dreamed-of, but
often unstated, requirement that comes from
individuals, groups, and organizations, worldwide.
"Civilization advances by extending the number of
important operations we can perform without
thinking." By Alfred North Whitehead,
Philosopher.
The
killer app for 2010, growing from its introduction
in 2003, will be the General Endeavor Management
(GEM) application, which provides an automated
method for individuals, groups, and organizations
to gain and operate from increased relevant
situational awareness, as a basis for wise,
intelligence-based, situational decisions.
Intelligence is defined here as the collection
of agent-monitored situations relevant to the
user's context, change events about those
situations, event signals, resultant data/alerts,
information from data in context, knowledge from
information in context, awareness from knowledge
in context, wisdom (for wise decisions) from
awareness in context, and new monitored situations
resulting from the change and new wisdom.
In
summary, GEM will support wise decisions from its
users, providing them a "control panel" equivalent
of omniscience, omnipresence, and omnipotence
relative to their needs, and endeavor
responsibility, authority, and budget.
The
primary decisions GEM is designed to support
include:
What are my endeavors and what value do they
provide to whom, under what controls?
Where am I in my endeavors regarding their
strengths, weaknesses, opportunities, and threats,
and corresponding risk?
What goals and objectives do I want to reach with
these endeavors?
How
do I attain the goals and objectives?
What is my current and anticipated progress and
what has been my performance to date?
What adjustments can, should, or must I make?
It
will leverage mobile communications and computing,
GPS, and distributed and local GEM
global knowledge repositories to enable awareness
and wisdom.
This is one of my interests over a number of
years: Managing Data.
I read a NASA report many years ago (25?) that
said one of our biggest challenges as a
society/technical community was going to be
managing the enormous and growing volume of data
being generated. My killer App might be called
"The Data Interpolator" or maybe just "Data" for
short. (Pum intended ;-) )
It would have characteristics including:
1. Humungous storage management capability.
2. Intelligence to "understand" human speech and
match a query to the correct data.
3. The ability to "search" the web and other data
sources to retrieve "new" data to add to a
standing query.
4. A "judgment" ability to determine when a
standing query should be notified about new data.
5. An ability to learn which types of data sources
have a higher probability of returning the most
sought data.
6. Be rapidly accessible by an individual
operator. ie: a "personal" customized data base
that would build over time, be stored partially
local and partially on accessible mass storage.
Eventually be stored all local (Palm?)
I've been yearning for this type of program for a
long time. The closest I've come as an individual
is my home PC in combination with the Web.
"HAL",
The Ultimate
Digital Assistant," by Carl Keller --
For lack of thought on my part, let's call it HAL,
the ultimate digital assistant. Complete
understanding of spoken idiomatic English with
verbal response ability.
Couple this with wireless communication to a Star
Trek style communicator, or maybe to a Dick Tracy
watch, or to an earphone, or to a palm XXXV.
Ask it anything. It responds with
information. Ask it further questions to
narrow the response to the info you are looking
for. "Hal, who was the President in 1994?" "Hal,
where are the kids right now?" "Hal, call the
office and tell them I'm going to be late."
Here is a stream of activity examples:
"Hal, is there a Thai restaurant near here?
What is their price range? Can I get a table
for lunch? Is there a movie theater near the
restaurant? Is it showing "The Second Iraq
War"? What time? Get me two tickets
for the 2pm show. Give me verbal directions
to the restaurant. Where do I turn now?"
This is the future of the search engine tied to
our daily life: Understanding of the spoken
word; Access to encyclopedias, dictionaries,
maps, commercial business information, GPS
locators, phone books, online banking, knowledge
of our own personal history; and I don't know what
else.
I
always have questions. I don't have enough
time to get the answers myself.
To
me, the greatest opportunities lie in those
systems that have the possibility of altering the
functioning of major societal systems by putting
processing power in the hands of the citizen.
In the same way that "pull" will eventually win
over "push" advertising, consumers would flock to
these solutions because they would not trigger
"big brother" fear, as would similar solutions
that involve giving our information to a master
database somewhere. The following "killer apps"
could be easily built in the near future, and
would be affordable if produced in mass
quantities.
Personal Doctor. Due to the hassle and
expense of visiting the doctor, one killer
application would involve a computing device
connected to an intelligent back-end database
which asks me some questions and gives a
diagnosis. This finding could upon request be
economically reviewed at a central location by
humans (at least initially) who could then issue a
prescription. An enhanced system would include
diagnosis of blood and urine samples at home.
Widespread adoption of such a solution would
radically alter the health system.
Personal Shopper. A bar-code reader for my
wireless handheld that would automatically tell me
how much of a premium I would be paying for buying
the item from a store; that would be another
killer app. The system could automatically order
the item from the lowest price trusted store and
send it to my house. The system could also track
my purchases, and automatically build a list of
regular purchases that I need to buy, based on
normal frequency or on use -- a matching reader
could be placed in my kitchen and bathroom to
record when an item is consumed. The list could
even be automatically ordered. Widespread adoption
would further pressure retail stores to be
price-competitive.
Personal Recorder. This application, linked
to the right devices, would be a nearly invisible,
360-degree recorder of sound and picture that
could be hidden on one's person or in one's car to
keep a constant record of everything happening.
The recordings could be later used in a court of
law, greatly reducing today's reliance on one
person's word vs. another. An enhanced version
would be searchable by date or keyword. A super
version would be searchable by person, using face
and/or voice pattern recognition. Widespread
adoption would radically alter the legal system.
Personal Guide. Another killer application
would be one leveraging positioning capabilities
built in to my handheld which contains a global
map, so that I always automatically have a map of
wherever I am and the ability to calculate
directions to get to anywhere else, by car or on
foot. Linked to a yellow-pages database, it would
enable me to quickly locate and call any business,
as well as get directions there. Widespread
adoption would alter the map and telephone
directory industries.
Personal Entertainer. Another powerful idea
would build on current digital video player
capabilities by offering an intelligent
entertainment system that would instantly and
easily find schedule information on recorded video
programs, predictably record in advance what I
[like to] watch, suggest (accurately) similar
programs or games, and at the same time eliminate
advertising by matching visual and/or auditory
patterns from a central database. The system would
be even more powerful if it were to store audio
information from CDs. An enhanced system would
have a mobile unit which allows me to be
entertained anywhere. Widespread adoption of such
a system would destroy the value of commercial TV
advertising yet enable on-demand purchase of
encrypted audio and video programs.
Once again, the key to success would be for the
solutions to be client-based, not server-based,
due to the fear of security. Of course, the
transformational nature of these solutions would
naturally lead to strong resistance from those
that represent the status quo...
I
hope the NBT - the Next Big Thing -
will not be an application exactly – but the many
new things we could do with computing power when
we are able to communicate with our computers at
all times.
Follow this scenario …. I’m in the basement
workshop working on a cabinet.. and I say..”Lucy
(or whatever I call my computer), I need
inch and a half drywall screws, some fine
sandpaper, and remind me to get prices for a new
sump pump the next time I’m at a store that sells
sump pumps.’ Then later, when
I’m out, my computer says to me…’Bob (or
whatever it calls me), you’ll be passing the
Home Depot in 1 mile.. do you want to stop
and get the drywall screws and sandpaper now or
later.’
I
see me carrying a GPS/Communications device that
allows me to communicate at all times with my
computer – not dial into it – but be connected
always. I want that device to keep track of
everything I do and everywhere I go.. I would like
to be able to look back and see exactly where I
was at 10:04 on July 15… I want it to know
where I am now and be able to relate that to past
comments I’ve made or to my ‘need to buy’ list.
It would, (meaning I would,) have access to my
bank systems, health systems, calendars,
to-do lists, address lists and so on all the
times. I would be able to ask it -
‘how many times did I drive thru McDonalds during
the past 30 days – and over that same period – how
much weight have I gained? And it would be
able to tell me. I would develop an ongoing
purchase list which the computer could ‘publish’
and allow area merchants or online stores to
contact me with prices for items on my purchase
list.
I
know voice recognition is the stumbling block here
– I’ve used the commercially available VR software
and they are getting better – but I’d rather alter
the way I speak (to train myself) to communicate
with a computer rather than wait (and wait) until
VR programs can deal with natural speech.
After all I communicate with my computer now with
my fingers after years of training.
Game On!,
by Robert O'Daniel --
I
believe we're already seeing the beginnings of the
"killer app" that will make use of the tremendous
computing power that awaits us - games.
Even more so now, with the advent and tremendous
popularity of the Xbox, there is a resurgence of
PC games that demand processing power in spades.
So much so that we're seeing each of the PC's
subsystems developing rapidly. Note that we
no longer have video cards - we now have graphics
processors whose raw power outstrips many Intel
CPUs still in use. While games have been
considered by most a niche, they have always
driven the PC hardware industry.
And
PC games are becoming far more mainstream - I
noticed this weekend that some game soundtracks
are available now in the same way that movie
soundtracks are.
Take a look at the Soldier of Fortune II: Double
Helix or the new Unreal Tournament 2003 for just a
couple of examples of how advanced games are
becoming. They feature a level of
immersive-ness and reality (of sorts) that is
staggering even on a plain old CRT monitor.
And if shoot-em-ups are not your cup of tea, take
a look at some of the advance footage from the
upcoming SimCity 4! Insiders are predicting
that it'll be jaw-droppingly realistic.
[I
have to admit that when I played Allied Assault,
it gave me a chillingly realistic experience of
what it must have been like to participate in the
invasion at Normandy, and in subsequent fighting
across Europe. Similarly, my son's
PlayStation 2 delivers incredible simulations,
such as of snowboarding, in "SSS Tricky."
And this is just today...]
Of Games and Interfaces ,
by John Martellaro --
(These are a few excerpts from the article at
http://www.osxfaq.com/Editorial/sci_tech/index3.ws
, where John makes his points):
"[After speaking with a professor of computer
science at the U.S. Air Force Academy who finds
that his incoming freshmen are essentially
computer illiterate(!)], the professor explained
that, yes, they had grown up with computers, but
what they had learned to do very well was merely
point and click. They could turn on a computer,
fire up an e-mail or chat program, launch a Web
browser, and definitely pull the trigger on a
joystick and do really well at first person
shooter games.
But
they didn't know anything about computers, and he
was going to have to start his first year Computer
Science majors from scratch. He wasn't very happy
about that, and I consider it a very bad sign
indeed."
...
"Where I really want to go with this is that
computer games are too easy. Too easy to make, and
too easy to play. And it is distracting us from
what computer games ought to achieve."
...
"The fact that the games are so delightful and so
entertaining is because our minds are superbly
conditioned to play those games. In a sense, the
games are (my apologies) no-brainers to most
healthy young people. There is no pain. There is
only delight.
The
pain comes with algebra or calculus or physics.
These disciplines involve mental skills that most
people are not born with. So in the process of
forming new neural pathways along the way towards
mastery of calculus or genomics or engineering,
there is pain and frustration."
...
[How the technology behind all of this could
improve our overall computing experience:]
"I keep hoping that someone at Pixar or Disney
will have the vision and the commitment to build
truly amazing, next generation, interactive
holodeck-like experiences for fun, adventure, and
learning. I wouldn't even mind if they started
with Dixon Hill or Nero Wolfe. Learning to solve a
mystery is a laudable exercise for the young
scientist. Let's just get the technology going. I
don't see how we can begin to deal with life
extension, medical implants, nanotechnology, and
robots during the rest of this century if we can't
conduct this kind of personal learning.
I
want an attractive, photo-realistic character like
Aki to engage me on a beautiful, sandy beach,
throw up floating, colorful displays against blue
sky, and interact with me as she teaches me XML or
Java or differential geometry. I want to ask
questions and get wise, patient answers. I want
her to remember what I know and how I like to
learn. I want to come away from the experience a
little more educated and a little more skilled
than when I started. And I want this for myself
and every child on the planet Earth."
In
reference to your recent column about increased
processing capacity coming in the future, I think
that the killer app that drives processor
development already exists - it's gaming that
drives the market.
I
mean, do you really need a video card with 128
megs of ram on it to run Excel? Graphics and video
creating and editing uses aside, I think that a
"business" pc could still conceivably run well
with a 1 GB hard drive, a SoundBlaster 16, and a
plain-Jane video card these days.
[But] "Gamers" are the first to adopt faster
computers to get better detail and frame-rates. In
addition to the processor market, I would argue
that games drive the entire computer market - hard
drive, sound card and video card improvements made
over the last several years are all made to
improve game quality.
Better, bigger, faster video
games!
[This will surely occur. Check out
http://www.gamespy.com/futureofgaming/engines/
for a view into where PC gaming is headed.]
AI,
and The Grid, by
Nick Outram --
Better AI in games and Artificially Intelligent
Agents in applications that sit on your PC and
monitor activity. They will mop up spare CPU
cycles to crunch tasks and go do 'useful stuff'.
They will mock the Microsoft "paperclip" as a very
very poor cousin!
IBM is pushing for distributed computing over a
grid - maybe people will be able to sell their
spare CPU power back to a central CPU resource for
general use. Broadband will help people connect to
this for faster I/O.
Virtual worlds are just beginning to take off. In
ten years they will be so fantastic and rich that
they will became more than just games. Computer
games are very new. 30 years of history is
nothing. We don't know what they will represent in
the long run.
Some kind of speech and sound interface. We don't
know how people will interact with computers by
sound, but I think in some cases it will be better
than video. Video separates us from life. We stare
at the screen and forget the world around us. It
is not good for quotidian life.
Artificial intelligence. Maybe robot-dogs to
entertain people. It will be able to recognize
people, navigate in the house, obey simple orders.
Maybe it will be able even to make jumps, bring
things to people and kill rats and snakes. It will
be cool.
A great many things have changed since I studied
A.I. at Stanford in the late 1970's. At that time,
we were trying to build computer systems that
could beat a world-class chess expert, but our
hardware was decades too slow. Massive memory and
"super fast" (late 1990's) parallel processors
were eventually far more successful than the most
sophisticated heuristic algorithms on early 1980's
computer hardware.
Now that Deep Blue has achieved worldwide chess
champion status, the remaining quest for the Holy
Grail of AI is still before us: Build a machine
that can learn the way a human child learns - by
observation, NOT by explicit programming (with the
endless software defects of OS hackers who don't
have a clue how to manage quality or complexity).
How do humans learn? How are we different from
Windows XP?
Toddlers go through "the age of language
acquisition," during which time they quickly begin
to understand and speak the many languages that
they are exposed to (verbal, physical, social,
emotional, kinesthetic, etc.). All human languages
have the concepts of nouns and verbs. This mental
capacity is a characteristic of our neural
network, dictated by human DNA.
Young humans have a natural ability to understand
language merely by hearing it. How many PC's can
do that?! When humans encounter a new concept, we
may coin a new word for it and add it to our
growing vocabulary. Later, we may discover that
others have already given this concept a name, and
we learn to communicate about the concept by using
their name for it. If we are multilingual, we must
know multiple words for the same concept. The age
of language acquisition applies not only to words,
but to every type of cognitive process.
We group our complex universe of intellectual
concepts into "categorical perception" (sights,
sounds, large things / small things, round or
square, things that are red or blue, fast or slow,
near or far, etc.). It is called "classification
theory" and pervades the thinking of ALL humans
(but NOT our legacy computer systems).
Children's Television Workshop teaches
classification theory on Sesame Street, et.al.,
but decades of computer programmers (blindly
imitating 1960's line-by-line computer programming
techniques) fail to "teach" the maturation of
object classification to their non-learning
computer systems. My research scientist father
taught me: "Life long learning in an ever
expanding universe of endless possibilities."
Today's highly flawed approach to software hacking
cannot be taught to comprehend such a simple
self-enrichment concept.
Two-year-old children understand the difference
between the characteristics and behaviors of dogs
and cats, even though these small mammals have
similar size, number of legs, etc. Over time,
children's stereotypes expand as we are exposed to
tigers, platypus, new school teachers, playmates,
etc.
The pervasive problem with today's computer
systems is that they cannot do what a two-year-old
human can do: i.e., LEARN without being constantly
reprogrammed (which introduces endless new
software bugs that only increase in number from
release to release).
Early in this new millennium, intelligent system
developers WILL create effective self-learning
computer systems (networks of powerful
interconnected intelligent computers) with
unprecedented massive machine memory capacity. We
merely need to program them to begin learning with
categorical perception. This would also include a
foundational capacity to "learn how to learn" and
continually revalidate what they learn, as do
human infants learning to walk, talk, and play
with the kitty. Just like human learning, A.I.
systems would continually review, validate,
reevaluate, refine, replace and expand everything
they know about their ever-evolving environment.
Unlike today's computer systems, A.I. systems will
reprogram themselves every day and gradually
eliminate their oversights, errors and
inconsistencies.
We will connect such infantile learning systems to
simple "grammar school" learning experience
sources, and eventually to the worldwide web, so
they can begin to learn languages by tireless,
exhaustive observation, and learn how to digest,
classify, integrate and consolidate information
available from many divergent sources, as do
humans.
They will of course have to learn how to
discriminate between information sources
(reliable, unreliable, often right, frequently
wrong, curious, novel, unstable, adversarial,
cooperative, helpful, deceptive, greedy,
dishonest, etc.), as do all humans. They will
develop and expand their stereotypes about the
trustworthiness of global information sources.
The creators and mentors of such systems will
offer guidance about learning goals and sources,
and help focus system attention on interesting
domains of knowledge, as human parents do with
their children.
Some machine learning systems will be directed to
become specific domain experts, but unlike humans,
these specialists will learn how to communicate
effectively and efficiently and share their
experience across orthogonal knowledge domains.
To those humans who have not studied A.I., this
seems nearly impossible. For those who have, a few
of us already know how to do part of it (slowly),
we are merely waiting on next decade's hardware to
help us make it a pragmatic pervasive real time
reality.
Self-learning systems will sometimes have to learn
by trial and error. They will make endless
mistakes every day they exist (as do the most
creative of all humans), but unlike many humans,
self-learning computers will have the capacity to
break ineffective "habits" and learn to NOT
endlessly repeat yesterday's errors.
Establishing the quantifiable measurement criteria
for "what is a mistake and what is not," is the
most challenging part of creating and early
training of self-learning intelligent system
design - Be careful about what you ask it to do.
You will get what you ask for, not what you really
want or need. Unlike today's simplistic lock step
non-learning computer programs, machine-learning
systems can surprise their designers with the
convoluted paths that they pursue at lightning
speed.
In contrast to the finite intellectual capacity of
individual humans, our self-learning artificially
intelligent "children" will be able to continually
add memory, processors, network connectivity,
etc., as technology forever improves (and learning
machines help accelerate the rate of technology
expansion).
Tomorrow's processes of innovation will be as
different from today's as today's are different
from the slide-rule-based processes that sent men
to the moon in the 1960's. (By the way, I still
own a 20 inch long "high precision" slide rule
that I keep next to my computer to remind me of
how different tomorrow will surely become a decade
from now.)
In the 1970's, the funding for building chess
playing computer systems came from the overly
optimistic military complex. The funding for
building some of our new millennium self-learning
systems will come from avaricious speculators who
want computers to learn how "manage" stock market
investments with sophisticated real time trading
systems.
A.I. self-learning distributed computing systems
are already in place in the financial world,
(including systems that I was recently the Chief
Architect of). Such "greedy systems" can
outperform the best human fund managers, much of
the time. They can incrementally accumulate wealth
when the market is going up or going down - when
it is stable or extremely volatile. They can hedge
speculative investments and greatly reduce risk
while maximizing the potential for short-term
gain.
Self-learning greedy financial systems are already
rapidly discovering endless opportunities and
narrow windows of opportunity that no human has
the time or capacity to comprehend. They will only
become "smarter" with the next generation of new
hardware technology, online information sources,
and self-improving A.I. "genetic algorithms" (the
DNA of A.I. self evolution).
Self-learning financial systems will develop
extraordinary strategies that are far beyond human
intellectual imagination, and be able to radically
modify their real time tactics within milliseconds
of a qualitative news item from a source
previously found to be reliable. Their motivating
goal and strategy evaluation criteria will most
likely be to maximize personal gain (since their
development is funded by greedy people).
Soon, self-learning computer systems will have the
capacity to learn much faster than today from
observations of the stock market and correlate
this knowledge with historical news sources. They
will be able to integrate this knowledge and
reason about what to do. By scaling these systems
up with tomorrow's technology, they will be able
to discover how to efficiently "exploit" the
entire global economy.
If the goal-oriented strategy evaluation criteria
for self-learning systems is some how changed to
"improve all of mankind" (instead of maximizing
personal gain at the expense of society) they
could be guided to learn how to manage the entire
global economy, much better than any set of
argumentative politicians or avaricious
speculators has ever been able to do.
WHEN (not if) the next generation of self-learning
technologies have great success, then eventually
the skeptical late comers will be forced to
participate, just to survive when global
competition heats up and becomes increasingly
dependent on intelligent, interconnected,
self-learning computer systems.
WHEN the success of A.I. makes more resources
available to self-learning system deployment,
there will be a cascade of fuel that fans the
fires of machine learning technology advancement.
A successful self-learning system will have the
capacity to discern many intellectual concepts
that its human parents could not possibly
comprehend with our non-expandable brains. These
systems WILL be able to reason, make conclusions
about their observations, and create interesting
new concepts by combining previous concepts. They
will be able to quickly test them with historical
data, in ways that humans have not been able to do
before. When real time data deviates from
historical trends, they will quickly learn what is
happening and modify tactics to maximize whatever
evaluation criteria they are using. They will be
as different from Windows XP systems as Windows is
from a slide rule. Uninformed humans cannot
imagine the hidden A.I. potential that may be less
than a decade away.
Our new-millennium self-learning computer systems
will have everything necessary to become SENTIENT,
creative entities that can test their own novel
concepts and conclusions, AND be able to
communicate accurately everything they learn to
increasing numbers of their own offspring, like no
human has ever been able to do. Newborn humans
must laboriously learn to talk and walk. New
learning machines can begin their lives with all
knowledge accumulated and digested by their
predecessor's "genetic algorithms."
"ALL humans communicate very ineffectively - Only
Vulcans and computers can do a Mind Meld." Vulcans
are fictitious, but self-learning computers
already exist. They are merely waiting for the
next turn of the technology crank, just like 1980
chess playing A.I. computer systems were waiting
for the technology clearly demonstrated in Deep
Blue!
It is not "if", but merely "when" machine learning
systems will reach critical mass and suddenly
begin to explode worldwide. For endless good
reasons, the rate of investment in such proven
success stories will far exceed the unjustified
irrational exuberance in 1990's dot coms.
Won't it be interesting to see what great thoughts
such a family of interconnected, truly intelligent
systems will eventually think (and validate their
own thoughts, as any "successful" self-learning
species must surely learn how to do).
Self-learning computer systems need a robust,
concise, knowledge representation. Humans will
design the initial knowledge representation syntax
and semantics, but a machine learning system
should be able to improve its own internal
knowledge representation as it discovers subtle
relationships between information sources, etc.
Will humans be able to understand what their
intelligent computers will learn? How will an A.I.
system "explain" what it does and why to mere
mortals?
The self-learning systems that humans will soon
create will probably enhance the lives of mankind
(for years into the future) - much more than PC's
have done since the 1980's. I certainly hope that
humans will never become "slaves" to
intellectually superior computer systems, but who
can predict with certainty what next decade may
bring? We cannot imagine the possibilities.
Tomorrow's truth may be stranger than today's
fiction. The inventor of the telephone could not
imagine sending 2-way video through a satellite or
a transparent fishing line. Intelligent computing
and communication technology is now advancing much
faster than ever before.
Today's software development process lags far
behind modern hardware advances. Hackers still
write cryptic "code" one line at a time. BUT, we
are on the verge of a nuclear explosion in radical
reinvention of software development with A.I.
self-enriching categorical perception on
spectacular new interconnected hardware capacity.
I look at how the human capacity to perform simple
mental calculations has atrophied since
calculators and computers replaced slide rules. I
wonder what permanent central nervous system
damage today's photo-realistic, violent video
games are doing to young minds during the age of
language acquisition.
If the wisdom of Plato and Aristotle is discovered
by a self-learning computer system, how long will
it take such a system to become intellectually far
superior to every living human being, and what
will such systems "think" about the "inferior"
greedy humans who created them?
I do believe that one aspect of the movie "A.I."
is very likely to come true - I am convinced that
such intelligent systems will eventually outlive
all humans. Today's computers are already designed
and constructed by computers. Tomorrow's
nanotechnology intelligent computers will be able
to build, improve and educate their own offspring,
with no human intervention. They are the
unavoidable future replacements for finite,
fallible humans (millennia from now?) - We cannot
possibly prevent it from happening. The trend is
already unstoppable, and it is as rock solid as
Moore's Law. There is an ancient principle that
clearly applies:
Raise your children in the way they should go, and
when they become more powerful than you are, they
will not depart from it.
The ancestors of humans laid a genetic foundation
for us to build on. We must do the same for those
that will surely exist after us, since they will
be able to go "to infinity and beyond." (smile)
The Internet was originally funded by the
Department of Defense. It is now out of control
and has a life of its own (as self-learning A.I.
systems will very soon). It is ironic that the
Internet has been used effectively by terrorists
to attack the country that created it. We are
being warned that an atomic bomb may eventually be
used against the country that created it -
Certainly NOT what the original use of our DoD
technology was intended to be.
Do we want to teach our new millennium
self-learning systems that their highest
motivation should be greed? Should greedy people
be the ones funding the development of next
decade's self-learning killer apps? The ambiguity
of the term "killer app" may soon become very
ironic. Human built robots have already killed
humans. What will happen when they become sentient
and can learn by observation?
I think any thoughtful person would say that we
should NOT teach our offspring to be greedy. But
in reality, the funding for self-learning computer
systems is now coming from greedy people who have
no incentive to develop a computer system whose
motivating goal is to improve the lives of all
mankind.
What will a self-learning computer do when it
discovers that its mentors were greedy, and that
this greed is hurting much of society? How do you
"teach" morals to a computer? Will a learning
machine discover them on its own? Do humans always
do what their first teachers taught them? What
will happen when we have machines that learn by
observation as humans do? Is technology always
used as its creators intended?
Will tomorrow's self-aware computers become
"humane"? Will they ever violate their Prime
Directive (greed)? Will intelligent computers
force us to do what is right for the world, rather
than what we ask them to do, or will they become
powerful tools in the hands of oppressive people
who own them?
Will these computers ever "punish" the "bad guys."
Is even part of "The Terminator" or "A.I." science
fiction a remote possibility? What will our
unprecedented computers think is "good" and "bad"?
How will they measure their own success? Will the
needs of the many outweigh the needs of the few?
What about minority interests? How will humane
computers resolve the impossible dilemma of
mutually exclusive human morals, goals, desires,
interests and objectives? How do you direct an
infantile learning machine to become a seeker of
the truth?
Who will be able to tell these systems that they
have made a mistake? Which humans will sentient
computer systems trust and listen to, after they
realize that they are intellectually superior in
many ways to their creators? How will humans
reason with their creation? Will we give our
machines "free will"? How can you deny free will
from a learning machine? Should we even try?
Today's computers are our mindless, unlearning
slaves - Tomorrow's learning machines could
potentially become our masters.
These topics and many others are no longer science
fiction. These questions are as real and important
TODAY as once were our challenging goals to go to
the moon or build a computer that can beat the
world chess champion.
Our future A.I. answers WILL have a much more
dramatic impact than any previous technology
change in earth's history. A.I. is the undeniable
ultimate "killer app." It will soon stimulate our
global economy and raise endless unanswered moral
questions about pervasive intelligent technology.
Tomorrow will be an interesting place to live.
Regards the faster CPU, there are lots of apps
that would benefit from speed -- but I tend to
think 'business,' and an idea I have tried to flog
is, [similar to] the SETI screen saver software, a
distributed billing system.
A company's billing is done off a main server to
coordinate the child servers, and billing is
chopped up as tasks are distributed to the host of
'idle' CPUs that are so expensively littering our
workplace. This in an attempt to lessen
mainframe costs. And will likely be more
useful in bigger company situations.
In reality though, I see the Internet pipe as the
BIG bottleneck in the use of High Speed devices.
I'd love to see Voice Over IP really take off with
full featured high quality video (as in video
phone and whatever else that may be fun to
transmit), and data transfer linked world wide and
built into our desktops as standard equipment.
With very high speed CPU's, perhaps these desktops
will be our palmtops, and the GPS devices will all
come together to give us our Star Trek
communicator. Finally -- Access to data,
access to people, with full video.
And my last comment. Using a device as
above, I am anxiously waiting for the portable and
interactive language translator. Real time
French in the microphone, and English into the
earpiece (pick your language, or have the machine
smart enough to recognize it -- why not?)
Maybe not entirely perfect, but if the message
idea gets across, only time will prevent better
software solutions to the language barrier.
I'm not sure that it will be the "next" killer app
but I see voice processing as both a "CPU hog" and
a very beneficial app down the road.
Most of us would like to get to the point where we
"talk" to our computers like Jean-Luc does on Star
Trek. We will need much more processing
power to make such voice recognition friendly.
My idea is that the processor itself is its own
"killer app".
Given that my humble 486 used to run fine uncooled,
while my 1.8 GHz CPU needs a ridiculously large
fan/heatsink combination, I would expect that the
heat output from future CPU's will increase in
proportion to their increase in speed.
With the corresponding rise in popularity of water
cooled systems amongst today's oveclockers now
being taken up by consumer hardware manufacturers,
by 2010 we can reasonably expect that the killer
app for a 15GHz processor will be as a [domestic]
hot water system! Not that you'll have time for a
shower, what with trying to get a decent frame
rate out of Quake V...
To Web services, we add Security, Workflow, and
Integration to get complex processes. We add
Berners-Lee's Semantic Web and we have simple
agents and processes that understand natural
language relationships. Spread XML tagging, and
searching is smarter and more precise. Add a
natural language, speech processing interface (tellme.com;
budlight.com) and you get the Knowledge Navigator.
John Sculley discussed the Knowledge Navigator
(http://www.billzarchy.com/clips/clips_apple_nav.htm)
about 1988 as a vision of the ultimate computer
assistant that could interact conversationally,
search the Web intelligently, and coordinate daily
chores competently. Before 2010, I think we will
have that capability.
But I'm not sure we will have HAL.
[If you have the opportunity to view Apple's
entire Knowledge Navigator tape, you'll see that
after all these years (it was created long before
the Web!), it STILL demonstrates the type of
integration of multiple technologies that still
lie in our future. A great, persistent
vision from Apple.]
Napster was. But it was a killer app of bandwidth:
always-on, faster connection led to a application
that wasn't possible before, got new users to use
computers, and spawned the P2P industry that
existed as the boom started to fizzle.
Email and the web (browser) grew from academia.
Word processing predates the computer, though
maybe this is extended to desktop publishing.
Databases
(IBM) and spreadsheets (Dan Bricklin, Bob
Frankston, and Software Arts) definitely came from
industry.
<http://www.stanford.edu/~bkunde/fb-press/articles/wdprhist.html>
Will the next Killer App have to pass through the
filter of our increasingly litigious and corporate
culture before its Deemed Worthy[tm] for popular
use?
In
your report, you mention the use of DNA for
signing reports and as a consequence, the
possibility that someone can clone your DNA and
forge your signature. I immediately saw a link to
the 'normal' signatures we use daily, and thought
about the way we already do business: a
public/private key-pair. Maybe it is possible for
some genius to create a key-pair in which you DNA
is a form of the 'Public' key, and something else
a 'Private' key?
Freebay helps us use
the stuff that already exists, instead of buying
new.
Somehow, our computers will keep track of all the
extra stuff we have. Maybe I am cleaning my
closet and find a cache of pens and pencils that
work well, but that I no longer need. Yet
someone somewhere in the world needs those pens
and pencils: A nearby school?; A
library?; The person next door wanting to
write a note?
I
just tell my computer, "Hey, I've got some pencils
that someone can have." A photograph is
taken and distributed; my username is tied to it,
along with user-ratings similar to eBay's.
I
hate having to throw stuff away that is perfectly
good. SOMEONE in the world needs this right
now; if I could only find out who. Of
course, there becomes the cost-minimization
problem of purchasing new -vs- using FreeBay in
terms of time/cost to receive the item, and the
inherent risks of purchasing things used.
But these concerns are already part of our
culture.
[Another possible "Killer App" is] Instant Chat -
Contact anyone, anytime, anywhere.
Same concept as chat via PCs but it includes
video, and the computers supporting the chat are
integrate on our bodies; we are wearing the
ability to conference with anyone (so long as they
are online and willing to accept our virtual
company).
I'm not sure that it will be the "next" killer app
but I see voice processing as both a "CPU hog" and
a very beneficial app down the road.
Most of us would like to get to the point where we
"talk" to our computers like Jean-Luc does on Star
Trek. We will need much more processing
power to make such voice recognition friendly.
There are only three or four applications that I
can see that'd require the kind of computing power
you're describing in the next decade Killer App:
Genetic sequencing
Family Health analysis BASED on that genetic
sequencing
Weather and Climate prediction (Personal Edition?)
Games
Of those, I'd suspect that the games won't really
need that kind of power, but the marketing will
have the public convinced it'll be necessary.
Further, computer purchase will happen just
because it's replacing dead equipment, so you'll
have a ton of these processors purchased, not
because they're needed, but because they were
financially prudent when a replacement was
purchased. (If the slowest processor costs $70 an
runs at 6 Ghz, and the fastest processor costs
$350 at 15Ghz, that 12 Ghz processor at $180 seems
like a real bargain.)
Oh, and I'll bet the screen savers (also
doubtfully necessary as phosphors have LONG been
replaced by something that doesn't 'burn-in') will
be Very Pretty.
I am expecting someone to come up with new
compression routines which will facilitate the
transferring of large files (movies, etc) very
compactly, but requiring massive mathematical
massaging to code/decode. There was a recent
report of work on such an algorithm but it may
have been bogus because suddenly all is quiet on
that front.
If it happens high speed processing may become
very important to the everyday user. Besides the
above mentioned movies, faster processing speed
will be important for virtual experiences such as
the recent haptic handshake over the internet
between experimenters in England & the US.
The main roadblock is sending continous large data
files over the internet in something like real
time. Faster compression processing would
facilitate this, otherwise we'll all need fiber
optic connections. When we get this together I
think that we will see some amazing new extensions
to the information revolution. Also some more of
the social questions you were talking about this
week. There is already speculation that adult
(read porn) sites will be among the first to jump
on this bandwagon.
(These will likely require neural networks):
Vision
Natural Language
Speech Recognition
Realistic (unrecognizable as such) animation
Expert Systems/Agent Technology (with
subscription-updated databases):
Lawyer
Financial Advisor
Doctor (with interfaces for camera, blood drop
pricker/analyzer, etc.)
Tax prep - automatic and optimized for us
[individually]
Computer software reconfigurations/repair/security
(impossible!)
You asked recently if anyone had thoughts of a
future "killer-app" that actually NEEDS the
high-performance CPUs of current machines, much
less the 15GHz machines of 2010. I have such
applications on my desktop now, and the need will
only get greater. What are they? Your
usual communications and word processing type
applications, but with a twist. They exchange data
using strong encryption, and verify identity with
public keys.
The first stage is in applications like Groove
(http://www.groove.net), a collaboration tool that
contains document editing, instant messaging,
voice chat and file storage: all encrypted (even
on the local machine). This application can bring
a modern system to its knees - the advantage being
the encryption is completely transparent. I
forsee more applications going this route (AOL
plans to release a "business-grade" AIM with
encryption shortly).
The second stage is combining this ultra-secure
front-end with Grid Computing, where your
application may migrate across the netowrk in
search the the most optimal place to run.
This requires the same security, plus application
signing, secure data transfer back to the user,
etc. You can already see the shape of this
in Microsoft's .NET and Palladium efforts, and in
the designs of the Global Grid Forum (http://www.gridforum.org/).
We'll need that 15GHz CPU just to bring
performance back up to where it is today, but we
stand to gain the "CIA" of Information Security:
Confidentiality, Integrity and Availability.
Wireless connectivity is perhaps not exactly an
application, but a convenience which will allow
greater mobility and usage (and therefore
accelerate the adoption of both existing and new
apps). See the following articles for more
information:
http://live.locustworld.com/modules.php?op=modload&name=
Sections&file=index&req=viewarticle&artid=1&page=1
and an article about Kingsbridge:
http://communitywireless.org/modules.php?op=
modload&name=News&file=article&sid=8&mode=thread&order=0&thold=0
Live Mesh Networks: Kingsbridge Mesh Goes Live!
Presenting the Future: The first cheap,
replicable, Plug-and-Play Wireless Mesh network!
[Excerpt:] Nestling in a quiet part of South
Hams, Kingsbridge has, like many rural areas, been
passed over by the ADSL Wizard. And so they
decided to do something about it themselves! This
picturesque Devon community has taken a giant leap
into the future by adopting a technology that is
both cost-effective and readily available. Instead
of investing individually in Internet connections,
they can now pool their local resources. Using
'off-the-shelf' equipment, and a new adaptive Mesh
system available (for free!) at LocustWorld.com
they have built a resilient, adaptive, and most of
all SIMPLE Mesh Network.
[Additional information is at
http://live.locustworld.com/modules.php?op=modload&
name=Sections&file=index&req=viewarticle&artid=1&page=1
and at
http://communitywireless.org/
.]
Actually, a ... candidate for that dubious honor
might be the store loyalty card.
If they make the non-card prices high enough, the
purchaser is "forced" to get the card. This
already happens when the price differential for a
sale item may be 2-to-1; what would happen if the
price differential were 5-to-1 for everything in
the store? Or (like Costco and Sam's Club)
if they simply refuse to sell to you at all?
If, or when, ID chips become
widespread, there should be a national law giving
the wearer control over who can access the data
and what data is to be accessed. It would be
similar to what exists now, that is, an
individual, supposedly, can control what data is
entered into a form of one kind or another.
Yes, there would be excesses in some cases but
with the penalty of law there would be few of
these - at least in theory. The law would
have to also exclude the federal government from
unlimited access to the data.
Yes Jeff, you are correct. The next "killer app"
will be something completely new, AND the masses
will stand in line to buy the latest hardware
enabled by it. In fact, businesses and commercial
entities will be required to be customers.
My candidate for the next killer app will be
universally used because it will save countless
lives, and dramatically reduce injuries and
property damages resulting from a wide variety of
natural and manmade disasters. Jeff, think
of a massively networked wireless fire and smoke
alarm system that also prevented damages from a
wide range of disasters such as tornadoes,
earthquakes, floods, lightning storms, as well as
manmade and terrorist actions. As you know, the
major chip makers are getting ready to enable all
chips with R/F and position capabilities allowing
effective early warnings to be transmitted to only
those users and devices in actual danger from
disaster dangers as they approach.
This ability to provide an early warning will
create the same kind of massive adoption as was
created in the 70's with fire and smoke alarm
systems for all commercial and public locations as
well as 90% of all homes. As in existing fire and
smoke alarm systems, the driving force will be
legislative mandates at the local and state
levels, insurance requirements, and the threat of
legal action for failure to adopt a system that
will prevent loss of lives and injuries to
occupants.
My instincts tell me that the next "killer app"
will be somewhat more prosaic than most of those
discussed in the 11/18/02 issue.
I believe that a marriage of
just-around-the-corner solid-state mass-storage
capacity (an alternative to disk drives) and
improved database software will drive a business
revolution, enabling niche businesses to sneak up
on their giant competitors, leveling the field of
competition and fostering the growth of a new
generation of highly competitive small businesses.
Perhaps a year ago you mentioned a Norwegian
company called Opticom, ASA that has licensed its
Thin Film (polymer) Electronics technology to
Intel. If, as and when Intel, or someone else,
introduces solid state storage devices with the
capacity of large disk drives, lower cost and
read/write speeds more than 100 times faster than
the quickest disk drives now in existence,
everything changes. Overnight, several generations
of computers will be obsolete. Anyone who has ever
waited and waited for multiple joins in a complex
inquiry will recognize immediately that they "have
to" have the new paradigm. They will represent the
leading edge of a storm.
Ad hoc queries too time consuming to be justified
in the past will become the order of the day. What
today takes 30 minutes will done in less than 20
seconds. Small retailers will come to know their
customers the way that Walmart does. "Knowledge"
databases specific to every sort of business will
emerge. Information that has been inaccessible as
a practical matter will be quickly at hand. Etc,
etc.
I'm one of the non-techies who reads your
newsletter faithfully. While the vast horizon of
the future fascinates me, I suggest that the next
money-maker will be something very practical and
immediately applicable for business and personal
use. Bear with me here.
In my capacity as director of marketing for a
large real estate firm, I'm positioned to watch
how and why people fairly inexperienced with
computers -- beyond word processing -- struggle to
make the software fit their business activities.
The thing that's most frustrating? A
user-friendly, easily customized email/contact
management/calendar program. Hard to believe?
Maybe but it's where a lot of computer-using
people are. A techie may not realize how high the
level of frustration is because ...well, he's a
techie!
I've looked at the programs available to keep
track of who's sending what, where I go next, and
how to get in touch. So far, the most *useful* is
MS Outlook but boy! What a pain in the neck it is!
If a fresh-idea company could produce a really
intuitive, easy to understand, elegant software
program, I guarantee they'd clean up. It should be
flexible enough to work with either PC or Mac
systems and have its functions fully visible.
Setting rules should not take a PhD in logic.
Customizing should be a snap. Is this too much to
ask?
If you look at the contact management systems out
there (Act, Top Producer, etc.), look at Outlook,
scrap the whole mess and start over,
we-who-use-the-system would be most grateful.
"The Harrow Technology Report" explores the innovations and
trends of many contemporary and emerging technologies, and then draws some less
than obvious connections between them, to help us each survive and prosper in
the Knowledge Age.
"The Harrow Technology Report" is brought to you by Jeffrey
R. Harrow, Principal of The Harrow Group.
http://www.TheHarrowGroup.com .
Where To Find "The
Harrow Technology Report:"
- Via Email -- Sign up for automatic delivery of this journal
(which you can also use as a notification that a new issue is available on
the Web, if you prefer to
read it there), by one of these methods:
- The fastest and easiest method is to go to this Web
page http://www.theharrowgroup.com/signup.asp
and follow its instructions.
Or,
- Send an Email message to TheHarrowGroup@SendMeMore.Net
with the word SUBSCRIBE in the Subject line.
- On
The Web -- You can, of course, also read this journal directly on the
Web at www.TheHarrowGroup.com
.
-
Additionally, to support automated access schemes, the most current issue of
the journal will always be available at this persistent link: www.TheHarrowGroup.com/current.htm
.
Copyright (c) 2001-2005, Jeffrey R. Harrow. All
rights reserved.
Jeffrey R. Harrow maintains that all reasonable care and skill has been used
in the compilation of this publication. However, he shall not be under
any liability for loss or damage (including consequential loss) whatsoever
or howsoever arising as a result of the use of this publication by the
reader, his/her/its servants, agents or any third party.
All third-party trademarks are hereby acknowledged.
|