The internet is the tip of an iceberg.  With perhaps 50m
people online now, and 500m anticipated in 5 years, it's
a world force growing with great momentum.  But for nearly
all users, the paradigm is a modest point-and-click: move a
mouse, point at part of a picture, and get another picture back.
I leave it to others to discuss the scaling of this new mass
medium, and the ensuing challenges (growing bandwidth,
dropping price of access). In this brief white paper I'd like to
focus on missing pieces: not simply "anticipated end-user
devices," but the not-so-anticipated devices that could really
have broad-based societal impact.

In particular, what happens when digital technologies collide
with ordinary objects?  What happens when everyday things
interact with networked links?  How will your jogging shoes
connect to the digital fabric, exactly; and when they do, how
might medical care be turned inside out as a result?  The most
interesting evolutionary infrastructural growth may occur by
inventing computers that aren't like PC's; by networking people,
places and things that have never been networked before; and
by focusing on new ways to link ordinary artifacts in ways that
might really add value to day to day lives.

The internet these days is largely addressing people through
email and web pages.  But people are outnumbered by things,
and it makes a great deal of sense to consider how those things
might interconnect in the future.

A future requirement will be that each person carry thousands
of networked devices -- not just phones and pagers, but sensors
and links woven into clothing, printed into the paper of books,
molded into fillings or pacemakers.  Prove this to yourself.
Look at the density of networked nodes:

How long will it be before the 10,000 things in your office are all
connected, somehow, to the network?  The infrastructure that
supports this will have characteristics that build on, but are very
different from, the building blocks of the internet.  Among the
requirements: new materials that can support forms of network
connection (printable radio tags, sensors, etc); network protocols
to track objects (beyond IP addresses, what will be the naming
scheme to track trillions of things in the network?); and network
meshes that organize themselves (you don't want to put on your
pants in the morning only to have to debug a network clash with
your shoes).

It is instructive to consider a few application domains.  I will
touch on three that have significant industrial and quality-of-life
impact: toys, health care, and the environment.


 

1. Toys of Tomorrow

Here is a very simple fact of life: when people play, they have
their best ideas, make their best friends, and learn their best
lessons. Outdoor games bring exercise for free.  Most adults
would be better  off if they played more, but "adult toys" sadly
lost their innocence and wonder a long time ago.  In retrospect,
a vast part of the computer and communications infrastructure
was bootstrapped through so-called  "office automation," and
enriched by university imaginations.  Well, it's said that play
is the work of children. What's going to happen when we take
the tools of their trade seriously?

Toys currently are not networked.
Video games and the latest "Barney" doll
notwithstanding, there are no components
and no agreement for networking toys.
They will be.  With a 30-cent radio tag and
a $5 tag reader embedded in every internet node,
every toy can trigger network responses.
With new infrared and wireless links, toys
can begin to interact in new ways.

Remember MIDI?  In some ways, the world's worst standard:
it was a hardware and software protocol for connecting music
instruments to PC's.  But it caught on, became the foundation
of a multibillion dollar digital music instrument business
(mostly based in Japan), and it's thought that Apple was
doing as much as $250m/y in ancillary music business
during their heyday, simply because people bought Macintosh
computers to make garage music.  MIDI actually played a
key role in energizing the personal computer industry.
There is nothing like "MIDI for Toys" yet, but a new standards
group is already building momentum for this.

Toys are well matched to digital evolution.  Put a chip in a
refrigerator, and you may have a smarter fridge, but it's not
going to budge for 20 years.  Toys by contrast pile into the
home every Christmas, every Chanukah, on birthdays, and on
lots of days in between.  The toy industry has a furiously
inventive pace -- 75% of toys are new every year.

Mattel knows this.  The latest generation Barbie dolls
connect to the internet.  Kids can use the Web to design
their own Barbie dolls -- the factory custom builds one,
to the child's specifications.  The Barbie camera (not yet
released) is a $65 digital camera that a child can use to
make snapshots, annotate diaries, even make animations.
The "tamagotchi" trend (a multibillion dollar new category
product for Bandai) points the way towards embedding
interesting computing in extremely inexpensive toys.
This year, Mattel will release a $10 toy "cyber" car, like
a  "Hot Wheels" toy but with an embedded display and
sensors. The car bleeps when it needs attention -- fuel,
tire changes, a pit stop.  Sensors track collisions so
the display can give you the body shop damage report.

Meanwhile, the computer industry is struggling to determine
whether a computer should cost $2,000 or $200.  A toy
needs to cost no more than $20.  (And it also needs to
be durable, fun, interesting, capable of surviving baby drool
and sibling fights: when Fischer-Price makes an email system,
it's not going to resemble the gunk we use today).  At $20
retail, a toy can contain about $5 in technology.  A great
deal of invention and creativity is be required to cast digital
technology in this form.  Enormous design ingenuity will be
needed, too, because poor use of technology can really foul
up play patterns, but when well used, smarter toys could be
profound amplifiers for a child's imagination.

Like cars, appliances, home security systems, and other areas,
development of infrastructure for toys will have a big impact.
It's essentially about creating a path along which some of the
most powerful technology ever devised will go right into the
hands of the future.  But the current business climate is
backbiting, with little cooperation among industry partners.
There are very few research efforts aimed at addressing this
(http://tot.media.mit.edu is one). And relatively few industrialists
who take toys and play seriously.
 

 

2. Health Care, Diet, and Sports

Consider: a runner runs the Boston marathon.  But the runner
is carrying a 1-pound belt containing a handful of health sensors.


Media Lab scientist Brad Geilfuss testing MarathonMan data pack.

Every second, the sensors record heartrate, footsteps, GPS
position, and body temperature.  (The temperature is taken
on the inside: the runner swallows a miniature radio thermometer).
These vitals are logged every second, and transmitted to the
internet. At any instant, friends and family can see where the
runner is, and how his body is doing: a sort of "black box"
crash recorder for personal use, but linked live, into the net.

That experiment was conducted in April 1997 and replicated
several times (with marathons in New York and San Francisco,
and with Army tests: see http://ttt.www.media.mit.edu/pia/marathonman).

A similar experiment involved embedded sensors in an
extravagent $500,000 diamond and ruby brooch, designed
by MIT researchers in partnership with Harry Winston:
 

Media Lab / Harry Winston "Heart Throb" brooch.

The rubies glow with every beat of the wearer's heart.
Small sensors in a brassiere track heartrate, and miniature
radios in the pin and in an accessory clutch purse drive the
rubies, and transmit the pulse on to the internet.

The implications are vast.  Put on your shoes in the morning,
and every day they know a little more about you -- much
more  than any doctor has ever known.  There is little question
that an intelligent toilet that made a mini medical analysis
on every deposit would have an enormous impact on day
to day medical care.

In 1868, Karl Wunderlich studied 25,000 hospital patients and
reduced their vital data to a single, standard chart: temperature,
pulse, and respiration were recorded on a clipboard at the foot
of the bed.  After antiseptics, this may have been the single
biggest improvement in quality control and clinical understanding
in hospital care.

Suppose your bathroom scale zapped your weight, every day,
into a medical file in your household "home page."  Twice a day
while brushing your teeth, the Braun electric toothbrush sends
a little infrared message bouncing off the bathroom mirror, as
your temperature and oral state are sampled.  Your doctor has
privileged access to your vital statistics log.

These sorts of technologies don't seem to be farfetched.  And
it is hard to tell how large the financial benefit might be, although
it is likely to be immense: such infrastructure will enable a
transparent, continuous daily personal medical chart.  That is
probably pounds of prevention, worth tons of cure in the hospital
system.  But today, there are virtually no networked medical
peripherals, and those that do exist have extremely limited utility.

Sports are the fun side of health.  Ambient technologies have
turned Americans into spuds: sedentary lives, spent in front
of telephones and computers, sitting in cars, occasionally
dashing for planes.  Sports might be made more fun, and
more insightful or inviting for amateurs, with a little technology.
A student at MIT recently invented "dancing shoes" -- shoes
that teach you how to dance.  Begin dancing, and a computer
listening to your footfalls through a wireless link plays music
to match.  Change to a waltz, and it plays a waltz.  What
may seem silly at first could be a "killer app" in the right
context ("karaoke for your feet.")

It is possible to build shoes that not only know a considerable
amount about you (pulse, pressure, stride habits), but can
begin to apply that information in new ways.  Go for a jog,
and the next time you stand in front of a screen, the shoes
pop up a map of your route and performance.  Fail to jog
often enough, and the shoes nag you by email. A reasonably
intelligent pair of shoes might teach you to dance like Fred Astaire.
 

prototype Media Lab dancing shoes
Our current sports system deteriorates rapidly from top down.
It's said that Michael Jordan earns more from Nike each year
than all the Nike factory workers in Vietnam.  Whether or not
that is true, and despite Nike's genuinely extraordinary efforts,
amateur sports are waning, in part because today's technologies
invite people to sit and watch, not go out and play.

The challenge here is in inventing nimble new ways to embed
sensing and communication in things like sports equipment
and apparel.  I would suggest that the active nature of sports,
the need for protection, the demands of outdoor play, are much
better drivers than Oscar de la Renta. Ultimately, this will help
build momentum for the general new area of well-integrated
wearable technology.

It's said that you're always happiest when something is entering
or leaving the body.  Yet most people are extraordinarily unaware
of their food consumption and what it means.  Despite standardized
nutrition labels, very few Americans know what 1000 calories and
50g of fat looks like on a plate, or means in their body.

Appliances, and common kitchen habits are in some ways the
toughest area to penetrate.  Talked about for years, the problem
is that price margins are brutal, rate of adoption or penetration is
slow, and most efforts to date are directed simply at convenience,
not at fun or family.  In the 1950's, the average American family
spent about 2 hours a day preparing a meal (shopping, cooking,
eating, cleaning, stashing leftovers).  Today the number is more
like 15 minutes.  Infomercials remind us that the device known
as a kitchen table is a wonderful medium for family communication.
A cornerstone of family life has crumbled.

A refrigerator could track food in and out, a kitchen counter really
could be a counter for calories, and a "kitchen sync" might be
able to hub the networking around the kitchen.  A combination of
food on the counter could fetch up recipes and riffs from Julia Child,
and in the long term provide valuable insight into consumption and
effects.  But left to their own devices, so to speak, the appliance
industry and white goods makers seem to be making poor progress.
Profit margins are brutal, and the digital industries have not provided
technology that can be incorporated in household appliances.
It's much talked about, but challenging because cost, pace of
change, and challenges of infrastructure are all extremely difficult.
And yet for something as basic to life as food, research on new
technologies that affect daily preparation and consumption of food
is remarkably scarce.

3. Environment

In early May, the American Everest Expedition will carry a
web camera and weather beacon to the summit of Mount
Everest.

MIT / Yale / NASA Everest team (May 1998)

This is a considerable technical challenge.  The parts may
fail, or weather may simply block the climb. If it succeeds, the
world will be able to watch the sunrise from the top of the world.
More important, if such a device can be made to work on Everest,
it could be built and placed almost anywhere -- at every UNESCO
world heritage site, from every street light and highway sign in
Los Angeles to watch traffic, or in parts of the environment that
are either too fragile or too hazardous for human visits.
MIT team with Mt. Everest satellite weather probes.
It is becoming clear that environmental awareness is growing
in the wake of what may be significant changes in the ecosystem.
But how well monitored is the world, really?  It is still remarkably
expensive to put a networked probe into the world.  Whether for
monitoring rainfall in the Gobi desert, adjusting precise control
over irrigation systems in farming, or just watering the lawn at
home automatically when it needs it (and when the internet-based
weather forecasts advise it), we are clearly hamstrung by the
lack of easy-to-use, automatically networked environmental
monitoring devices.  Consequently, we have a rather poor grasp
of the real behavior of the planet as a system.  And this comes
at a time when, unless there is a significant scientific breakthrough
or natural cataclysm, there will simply not be enough fresh water
for the people on earth in 50 years.

Investments in this area are small.  Technologies are specialized
for tiny niches.  The economic base (whether farmers or home growers)
does not seem to be yielding creative, broad-based solutions.
Advances tend to reduce biodiversity: most of the world's foodstock
now comes from about a dozen grains and animals, because
fertilizers and mass agriculture focus most efficiently on a
few species.  If we want to manage the environment sensibly,
we need to understand it accurately and control our input wisely.
To do this will require major new research and business thrusts.
 

 

4. Technology Investments.

The bulk of digital technologies are aimed at personal use
(computers, phones, cars, etc) with a large stream of rather
unsophisticated, disconnected controllers aimed at dumb appliances.
We are building billion-dollar silicon plants to produce memory
and cpu chips, launching satellites to better enable worldwide
telephony, and we have created an organic climate in the internet
that fosters a great deal of entrepreneurial activity.

But large scale economics and investments in digital infrastructures
are not yet aimed at some of the most fundamental aspects of life:
play and all the media of play; day to day fitness, medicine, and diet;
and the natural world around us.  It is likely that capitalistic systems,
which are inherently profit-maximizing and now are globally driving
most industries, are just not suited to finding breakthroughs in areas
that are not overtly or immediately profitable.