|
A Dialogue On Cycles, Part I and II
01/09/02 Part I - My friend and colleague Mark Stahlman recently gave a presentation at Internet World on a theory of his; that we fail to understand the successes and failures in the information industry companies because we misunderstand the underlying economics. We’ve been corresponding on the subject and I think you’ll be interested. Read Whole Story Part II - The Dialogue I started with Mark Stahlman on the cyclical nature of products stirred up some lively correspondence on the nature of markets, products, and cycles. We present some interaction between various correspondents here. Read Whole Story
My
friend and colleague Mark Stahlman recently gave a presentation at
Internet World on a theory of his; that we fail to understand the
successes and failures in the information industry companies
because we misunderstand the underlying economics. We’ve been corresponding on the subject and I think
you’ll be interested. Mark
is a fairly vocal analyst/writer.
He describes himself as the analyst who first recommended
Sun Micro and the banker who brought AOL public.
You might also remember him for the minor furor stirred up
by his "Why IBM Failed" article for the Harvard Business
Review in 1992/3. He
was a co-founder of the New York New Media Association and is a
frequent speaker. He
is currently raising a venture fund. Of course, you’re welcome to chime in. Mark is even fonder of debate than I, so this could get quite lively.
Dear
Amy: A
while back MIT economist Paul Krugman claimed in a NY Times column
titled “Chip of Fools” that economists simply can’t figure
out how Moore’s Law fits into economics. As a result, he claimed, we are endlessly making mistakes in
forecasts and being surprised when markets go haywire. But
that can’t be right, because you’re an economist and surely
you’ve figured this out! There
is another, perhaps more troublesome, version of confusion about
this critical issue floating around . . . more troublesome because
it’s being promoted by people who are well regarded in the
technology business. This
version is what I call the “humped exponential.”
Amazon CEO Jeff Bezos used it – to justify what he called
his nonstop, grinning optimism – at last year’s PC Forum.
And, various Morgan Stanley analysts (including Mary
Meeker) used it again just last month at the firm’s Technology
Summit. I’m
sure that you’ve seen it. It
looks like an exponential curve – rising faster and faster from
right to left – with an extra “hump” on top of it and it’s
supposed to be a chart of “revenues” over time. The
theme is that while things may get overheated from time-to-time
(i.e. the upside of the enthusiasm “hump”); the underlying
principle is an ever-rising baseline that will always reassert
itself whenever a temporary downturn (i.e. the downside of the
enthusiasm “hump”) takes over. So,
Bezos reasoned, Amazon just has to succeed. Since Moore’s
Law drives this rising baseline and since Amazon is obviously
following the curve, why worry about something as short-lived as a
temporary downturn? Don’t
worry . . . be happy! Self-serving? Sure. But,
that’s not the issue. The
problem is that Bezos (and others) are just making Krugman’s
point and adding to the confusion by blaming market
“irrationality” for these swings . . . and therefore
compounding the mistakes and often painful surprises down the
road. I
think it’s possible to shine some serious light on this matter
and to try to nail down some solid economic principles based on
the past 35+ years of Moore’s Law operating in real markets with
real economic consequences . . . and I suspect that you agree. The
research that I began on this topic in the late 1980’s pointed
to some harsh conclusions that have been born up with a vengeance
in the past few years. Briefly,
the computer/networking business is CYCLICAL but on a FLAT growth
baseline, not on an ever-rising one.
These cycles are driven by Moore’s Law, for sure . . .
but the ups and downs are specifically the expression of the
well-established “S-Curve” of market adoption of innovations
and not some sort of “irrational” waxing and waning of
techno-enthusiasm. What
we know from the study of “S-Curves” is that all
markets inevitably reach relative saturation and that the
high-growth phase is always in that period in between
early-adopters and the latecomers. This is Market Economics 101,
right? The
Internet (like all other computer markets) was no exception.
As an innovation the Internet began to pick up steam in the
early 1990’s, reached its highest growth phase in the mid/late
1990’s and – inevitably – hit relative market saturation by
the 1999/2000 period. All
the indicators were there. So
many people already had Internet access by 1999 that new user
statistics began to level off. Correspondingly, the sales of the equipment to operate the
Internet – Pentium-powered desktops, Cisco servers and Sun/IBM
servers – had to begin to decline.
Just had to . . . no matter what Moore’s Law kept on
improving chip economics. By
2000 the high-growth phase of the Internet was over and it will
never come back. The
cycle was coming to a close.
Just as all the other cycles before it.
The law of the “S-Curve” was taking its toll. And,
ignoring the consequences only compounded the pain. In
fact, my early research showed that mainframes, minicomputers and
standalone personal computers all followed the same
“S-Curve” pattern. Surprisingly
to me, these cycles all had more or less the same period of eight
years as well. Just like the Internet. I
could go on adding details (particularly on the crucial matter of
how you define these “markets”) but you get the picture and,
hey, you’re the economist so I want to hear what you’ve got to
say. Mark
Stahlman
Dear
Mark: They
don't call economics the dismal science for nothing.
Economics
is useful for thinking through theories and trends, but not very
useful for the kind of precision that venture capitalists and
CEO’s are looking for. That
is a fine point often lost on the clients who buy their services. I
think that it’s easy to see how we get to where Jeff Bezos, Mary
Meeker, and others find themselves.
They are simply taking a number of experiences they have
had and assuming what the underlying relationship might be.
In fact, the underlying relationship might be different –
or there might be no underlying relationship at all.
Those of us trained in economics are (or at least were)
taught to be careful about assuming cause and effect relationships
unless we could prove them. All
of these folks have spent their business life during the years of
the PC era, during which Moore’s rule prevailed, and during the
longest bull market in history.
No wonder they think in terms of rising lines!
All
of us were schooled in overlapping “S” curves representing
waves of new technologies (or new major products/categories of
products, based on enabling technologies).
Those S curves were typically drawn, by the way, to a flat
baseline. That was
the idea. New ideas
entered the market with nothing, crawled around on the bottom for
a bit, eventually went through a period of rapid growth, and then
tapered off as (a) markets were penetrated and new and (b) better
ideas came along to compete for customer attention. Note that (b)
was often at least as important as (a). There
has been an argument, from time to time (it is, in part, the basis
for the argument for why monopolists must be prevented from unfair
competition in other markets), that established vendors can enter
new markets more readily than new vendors since they already have
existing resources (cash, infrastructure, employees, markets,
customers, partners) that they can leverage into success in new
markets. In many
markets, it is easy to make the counter-argument that new vendors,
with a clean sheet of paper, also have advantages (no existing
products to hold them back, no existing cultural issues, no
competing with themselves, etc.).
New vendors are also lean, hungry, and very willing to work
long and hard. In
computers (Digital for minicomputers, Apple for PC’s, Sun for
workstations), in telecommunications (Cisco for routers, Nokia for
wireless telephony, AOL for Internet services), it is new vendors
and not established players who have led the way. Of
course, from time to time, smart management can “reinvent” a
large organization and refocus its resources and efforts.
Microsoft did that for the Internet (probably too late to
be really as important as it had hoped) and again, more recently,
in Web Services (perhaps in time).
IBM, with the entry of Lou Gerstner, reinvented itself into
a services business, judged by the complete solutions it and its
partners can provide – a very hard act to compete against.
Such over-arcing strategies may be able to bridge
technology cycles and outlast them. So
I would agree with you that it is “right” to assume that when
new markets are starting, they are starting from the ground up,
and not from where the last market stopped – usually. Amy
Wohl
(At
this point I sent Mark a copy of #65 of David Isenberg’s SMART
LETTER, a very good newsletter on telcos and networking.
In that issue, dated 1/3/02, David and some of his
colleagues discuss the problem of how to value, regulate, and
re-invest in communication infrastructure.
You can find David at www.isen.com.)
Amy: Which
argument (he means in the article)? The
depreciation of switches one? My
view is that telecom is NOT the same as computers/networking . . .
as indeed, automobiles, airlines, healthcare and umpteen other
businesses which have lotsa Moore's Law in them aren't the same
either. Telecom
needs JUNK BOND funding and GOVERNMENT permissions, to start with,
and that's not going to change.
Computer industry financing works differently. And
it's telling in all the stuff the SMART LETTER readers rumble on
about, there is no recognition that the INTERNET reached relative
market *saturation* in 2000 (and that this could have been
predicted in, oh say, 1989!) . . . which none of the CLEC's
apparently appreciated and thus, the "telecom
recession." Mark
Stahlman
Mark: I
want to have the debate in the letters column, so if we have it
here, expect to see me cutting and pasting. There
are no neat and tidy boxes with crisp dividing lines.
The Internet is how we connect computers and may ultimately
serve as part of a giant computing system.
And yet it is also a communications device which we connect
to (most of us) through telco services.
Telcos could not exist except through their dependence on
modern computing technology.
I
would suggest that although, indeed, their funding and regulation
have been very different the interesting arguments David
Isenberg's readers presented are germane. It is because telcos are being transformed by an underlying
technological change (that may cause the very notion of a telco to
become obsolete) that this is so important and interesting. Amy
Wohl
Amy: Hmmmm
. . . not sure what you mean, but if we wish to do ECONOMICS then
we will have to separate ELEPHANTS from EYEBALLS, right? <g> The
business models in telecom *are* changing, but not quickly enough
or far enough to make a CLEC into a Sun Micro . . . or to make the
sort of people who run (and finance) them into Scott McNealy’s
or Bill Gates, as far as I've observed. My
sense of the most important BLOCKAGE to a clear economic
"science" here is that everyone wants to believe that
they can be in whatever business they wish to -- as taught in
every B-School (and, ultimately, the reason for HBR's crisis over
my article) -- so that no one is willing to draw the necessary
LINES . . . and the result is "humped exponential"
pseudo-scientific MUSH. Mark
Stahlman
Dear
Mark: I
had no illusions that the telephone company would find itself run
by Scott McNealy. Rather, I was offering you David Isenberg's argument that
because it wasn't run by someone like Scott or Bill and because
the Cro-Magnons who do run these bureaucracies are using
accounting rules that guarantee that they are miscalculating the
value of their assets and the costs (lower) of their new,
clean-sheet competitors, they were very unlikely to be able to
compete. In
other words, I was suggesting that I didn't think the way we would
all be connecting to the Internet in a few years would be via Ma
Bell but rather through cable et al. And then we may start making the rest of our phone calls (the
ones that aren't going through our wireless phones) through this
same (or some other) non-telco network. I
think the lack of crispness isn't that we can't create an
appropriate business model each time but that the business models
wear out (you would argue that their limit might be tied to that
eight-year cycle) and the executives who run the companies fail to
understand that they need new, different, business models. They keep trying to recycle the ones they know already.
Boring. Futile. Sure to fail. What
people need to understand is: --
How do you tell which business model is right (for the business
and for the times)? --
How do you know when it's too late into a cycle for a new entrant
to successfully start into the game?
Is it after the halfway point?
Is there some recognizable milestone event that occurs?
Is it different for different cycles? --
How do you figure out what the next technology cycle will form
around? Is there a
period of time before it starts when there are multiple candidates
milling in the doorway, waiting for the selection process to
occur? How do you
know it's autonomous computing just getting starting rather than
further into a wireless networking cycle (or is that a sub-segment
of the Internet cycle)? --
And most interesting (to my mind), how do you know when it's time
to jump in? I've seen
as many players burned by trying too soon as too late.
In fact, my whole grad school course at U of P this term is
about that -- why some products fail and others, seemingly
identical or nearly identical succeed.
I maintain timing is an important component. Is it? Amy
Wohl
The
Dialogue I started with Mark Stahlman on the cyclical nature of
products stirred up some lively correspondence on the nature of
markets, products, and cycles.
We present some interaction between various correspondents
here. It’s very
long (these are passionate people), so if this isn’t your area
of interest, we’ll understand your skimming or skipping this.
From Bob Brown, ex CEO of market intelligence house Computer Intelligence/Business Intelligence/Harte-Hanks and now a strategic consultant: Amy: I
enjoyed Mark and your discussion on future demand scenarios.
Couple of additional points which support a stronger growth
picture than you’re outlining: McKinsey
presented a large study at Agenda which showed that only 6
industries gained high productivity improvement in the go-go years
96 – 99. The assumption is that the other industries are
poised to gain that higher productivity; hence the demand may be
sustained for years to come. Maybe it’s likely that the
others won’t create improvements as rapidly, but there’s a
good opportunity for most businesses. The
Harte-Hanks data supports an interesting demand snapshot.
It’s no secret that tech demand was way off in 2001. It
was 25-40% below last year (2000) in most segments except
government and education. However, looking deeper it appears
that business is still spending on crucial technological
differentiators. So, GM is spending on On-Star but cutting
like mad on infrastructure and applications for net flat spending.
So, irrespective of current gloom, we have a whole new area of
demand engendered by the web which is layered on top of the
historical demand areas. That’s hopeful if people can
afford all three. Regarding
telecom, my all time favorite article title was by Bill Gurley who
wrote “Backhoes Don’t Obey Moore’s Law”. His point
was that high bandwidth is crucial to some of these high growth
scenarios and that it is slowed by the physical reality of getting
homes and businesses connected properly. Robert
G. Brown
Bob: I am a
long-time debater in the productivity dialogues. I agree
that we have lots of headroom left there. The trick is
figuring out where it will come (and when).
I wouldn’t count on displaced spending necessarily
reappearing in a predictable way, though.
I think when it’s gone, it may be gone. Amy Wohl
Opinions
has heard from Darryl Carlton, Entrepreneur, before, on the
subject of business models. Amy: There
seems to be an "implied assumption" in all of these
discussions (yours and elsewhere) when discussing market forces
that suggests that "the market" is some OTHER group of
people. The
market makers of the internet bubble were the professional
investors, particularly the private equity and venture capital
investors, closely followed by the Investment Banks. Joe Public
invested because he thought that the professionals MUST know what
they were doing, and damn! they were making so much money he
wanted a piece of that as well. These
professional investors were driven by massive and unbridled greed
- they just had to follow Amazon, and were prepared to hype
anything so long as it was "name-your-segment" version
of Amazon. DogFood.com raised more than $35 million because it was
going to be the Amazon of the pet food business. The
investors were incapable of evaluating the market potential, they
were unwilling to put in the effort to understand the technology,
and they were unprepared and totally ill-equipped to manage the
businesses that they were investing in. As
a collective this group had the attention span of a gnat - and so
as soon as things got tough (i.e.: they had to build and deliver
product and run businesses) they panicked, pulled the plug on
their investments, and ran to the nearest excuse (blame the
entrepreneur, blame an overheated or irrational market). All of
these excuses are far too simple and bland. Companies
take a long time to build - the only real issue with internet
companies is that they should never have been brought into the
public market so early. Private Equity investors abrogated their
responsibility to fund businesses for the amount of time it takes
to make them successful. Private Equity tried to make a killing
off each of their investments by pushing them into the public
arena before the companies were ready, and before the markets were
ready. Greed kept them afloat for a short time - but it takes more
than greed to keep companies going for the long haul. All
of these discussions are still confusing the business markets with
financial markets - these are fundamentally different. Many
companies that would have made it in the business markets, failed
for lack of funding from the financial market - and many companies
that should never have been funded (lousy or non-existent business
models) made millions for the founders and venture investors, only
to fail in the real world when their funding ran out. The
internet is now in the Chasm - and this has two sides to it, the
downward slope has all of the established players desperately
trying to leap across to the upward slope. Their argument is their
established strength, their customer base and their credibility.
The upward slope has the upstarts with the new ideas, the new
technology, not enough customers and pitiful levels of funding. The
real growth of the internet is yet to come - the old players will
be replaced, radically new technologies will usher in a whole new
generation of growth. What we witnessed up to 2000 was just the
new market clearing its throat. Services will cause a massive
shift in behavior and wealth creation. Companies
will become successful as businesses, financial markets will
consequently be "less thrilling" but more predictable,
and values will rise over time to massive new heights. Darryl
Darryl: I
agree with some of the things you say, but not with others. Mark
Stahlman does make (perhaps not in what we've published so far) a
very specific difference between economic and financial markets,
similar to yours. The
Internet Bubble is, in some sense, almost a separate phenomenon
from the question of whether the Internet, like other
technologies, has an eight or so year growth cycle that has now
flattened out. It
exacerbated the problem, of course, that it (the bubble) overlaid
so sharply, the ending of the rapid growth period of the Net. Amy
Amy: With
reference to the rapid growth of the Net, why do you assume that
is over, just because the vapid greed has diminished. Perhaps now
we can grow the business end of the internet - its growth period
for USE has yet to begin in earnest
- and if there
is some degree of skepticism from the investing community, then
maybe it will grow properly and usefully, rather than being driven
by the particular whims and fancies of private equity investors. I
also disagree with the 8 year time frame - its more like 30 years,
with an 8 year period of peak activity which corresponds to the
Geoffrey Moore version of the technology adoption cycle, with
overlapping cycles in successive waves (ala Kondratiev)..........
at the point of intersection of old and new cycles we get a
chasm for both the existing and the new players. The old
technology will still continue with diminished importance for some
time to come, while the new technology grows in importance and
eventually to dominance. Darryl
Darryl: I
think we are differing not in what we think, but rather in how we
are defining "technology." You
are, of course, correct that the technologies don’t
“disappear” in eight years – we are talking about their
growth period, not their period of being available in the
marketplace. In
the "eight year theory," technologies are more narrowly
defined, so, for example, we don't talk about computers, but
separately about mainframes, minicomputers, and PC's, for example.
I think the tapering off of rapid Internet growth Mark is
referring to (which is measurable, by the way, and noticed by
many) is not THE INTERNET, but rather the World Wide Web and its
use as a growth engine for the development of community.
I suspect that the Business Internet, which is still in its
early stages, is a different phenomenon which may be tied to
different players. We
would agree, for instance, I think, that AOL, a very important
winner in the Internet round, is unlikely to be an important
player in the Business Internet round. I
think we're getting ready (very early stages) for a whole round of
Home- and consumer-related expansion but I'm not sure it is
entirely Internet-related. It
may be more related to removing the notion of
"computing" from the equation and allowing the
buyer/consumer to think of these devices as info-entertainment
appliances, much like TV's and DVD players.
In that, I have high hopes for Moxi, for instance, and low
ones for the theory of the universe that says consumers are
waiting to line up to buy Apple and Microsoft PC's which will be
the hub for their wireless home networks and new devices (Mira et
al). Amy
Mark
Stahlman, has, in the meantime, been getting copies of the
correspondence between me and Darryl and has been chiming in. Mark: Ubiquitous
computing - the next great wave of consumerism - I agree ...
It’s not a computer if the purchaser/user does not consider that
they are using a computer ---
if it is just a device that they use to complete a given
activity, then computing becomes ubiquitous and I like the idea of
separating business internet from the community internet .... and
it is true that the business internet is yet to take off. Darryl
Darryl: “Ubiquitous
computing - the next great wave of consumerism - I agree” Yup,
but first we have to solve the problem of computers that are all
too obviously computers . . . the crisis of complexity in which we
are now stuck. This
will NOT start with consumer products, it never does, but will
begin with much more expensive systems that blend in with our
environment. And,
that's what will be the focus of the SIXTH cycle. Mark
Mark: Yup
- I'm at
Tonga, palm trees, white sands :-) and it fits my new venture ---
Carlton Training Organization As
to the crisis of complexity ........ again the erudite Yup! is an
appropriate response here. I
would state that this is a software problem principally - while
hardware and networking will make great strides over the coming
years, the current infrastructure is perfectly adequate to support
new architectural models of software; they just ain’t being
built ---- and I have previously corresponded to Amy on my opinion
of MSFT's ability to Architect !@#$%$%^&& This
will take some radical thinking - and the preparedness to
experiment and fail on the path to magnificent success (and riches
the likes of which would make Larry Ellison green with envy) In
fact - the technology models are reasonably well known and
articulated today, they just don’t exist in commercial practice
'coz there is a risk inherent in change (downward slope of the
chasm versus the upwards slope) I
agree with you that this revolution will start with large scale
computing - my feeling 7 years ago when I started to build
software for this purpose was that it would start with the
Telco's, because they would evolve into service based organizations
and the new architecture demands "software as a service"
in the most extreme sense. I further postulated and have restated
many times, that ultimately banks and telcos would merge as their
businesses collide. My feeling has always been that the Banks
would end up the dominant partners in this arrangement because of
their management style, but who knows. Darryl
Amy: Regarding
Daryl's comments -- yes, I think that you're right that
definitions are crucial if we are going to try to make useful
distinctions in this economic analysis exercise. If
we think of the Internet as mostly Pentium-powered PCs connected
to a global network based on Cisco routers feeding into/off-of
Sun/IBM servers, then this market is largely finished.
It has reached the top of its "S-Curve" and this
is reflected in the sales of each of these components. If
we use Internet to mean something else -- like the total universe
of all TCP/IP equipped devices, for instance -- then there are
undoubtedly many markets that haven't even begun yet. However,
whatever you mean by the Internet, your definition can't solve the
fundamental problem of what is being called the crisis of
complexity. In fact, the broader the definition, the worse this crisis
becomes! So,
my best judgment is that, while the growth of the Internet
(broadly defined) will surely continue, what we is need something
much more like a basic breakthrough in platform design to drive
the growth (and value creation) in this next cycle.
Sure, some people went to cheaper machines in 1986 (i.e.
the parallel to today's fascination with "consumer
electronics"). But
back then, the long-term value was being built in networked
workstations -- which then gave us the Internet -- not in low-end
"one-trick-pony" standalone devices. Much
like those fascinating years of 1986 and 1994 when nothing seemed
very clear and the consensus was about to overturned, right now in
2002 (eight and sixteen years later), we are in need of some fresh
new, breakthrough thinking. Thanks
for providing a forum for discussing some of these questions! Mark
And
further . . . Amy: Let's
try to sharpen this up a bit, okay? There
are many different "markets" and "economics"
that have a tendency to get mixed up in these conversations.
My career has convinced me that making distinctions is
usually the most difficult (and often the most rewarding) of
exercises. So,
let's make some distinctions.
What are we (and what aren't we) talking about?
And why? We
are not discussing the financial market for technology related
securities. We are also not talking about the market for any
particular technology (such as TCP/IP or HTML or 1.3 micron wafers
or CAT 5 cabling) . . .per se. We
*are* trying to discuss the markets for applications driven
products which aggregate into sweeping collections of business
activities that can together drive billions of dollars of revenues
over a relatively short period of time. Before the term became
misused, we used to call this kind of major market phenomenon a
new "platform." Why
make these distinctions? Why
try to carve out whichever particular "platform" that
drives sales now . . . and in the future?
So that we can get ourselves to the right place at the
right time, of course! We
all suspect that the sales of technology-based products will -- in
aggregate -- continue to rise indefinitely.
So, this knowledge doesn't really inform us or help us make
decisions. We
also all know that securities markets will fluctuate based on the
instantaneous balance of fear and greed in the minds of the
participants. To be sure, we all try to second guess these investment
"mood" swings, but however good we might get at
understanding this psychology this knowledge also doesn't help us
very much to make fundamental technology decisions. What
we really want to know -- in a fundamental market
"economics" sense -- is how does this aggregate of the
sales of technology-driven products (and services) break down?
What are its components over time?
Where does the growth come from?
What will be the next "platform"?
When will all this take place and how do we participate n
this growth? My
Wall Street career forced me to try to answer these questions. My "niche" was to be a "fundamentals"
analyst. Mostly
unconcerned with day-to-day trading patterns (i.e. commissions) or
with investment banking fees, I was paid to try to figure out how
the computer business really worked.
I know, bizarre, but true. So,
I literally took all the available sales data (from IDG, Datapro,
company reports, Commerce Dept., you-name-it) and went through
year-by-year, company-by-company, and product-by-product for the
20 years from the mid-1960's to the mid-1980's.
I made distinctions based on the generally agreed upon
categories of mainframes, minicomputers and personal computers.
This was relatively easy at the time since these products
were largely stand-alone machines sold mostly by different
vendors. What
I discovered surprised me. The
source of growth shifted almost totally from one category of
machine to the next and the shape of these growth-contribution
curves was remarkably regular and periodic.
Then I met an IBM economist (he was probably the only
person on earth who had more data than I did at that time) and he
explained Pareto distributions and some of the other underlying
reasons for this phenomenon. So,
standalone computer sales moved from expensive to cheap over the
course of 20 years and it took three "cycles" for this
to happen. Each cycle lasted about the same time of eight years.
Each involved the mobilization of dozens of primary
suppliers and hundreds of supporting players. Tens of thousands of people were put to work to construct a
new "industry" with a new set of rules.
Digital Equipment and others emerged as a dominate player
in minis. Microsoft,
Compaq, Lotus and others appeared in the PC "Wave." So
what happened next? Even
cheaper standalone machines?
No it was evident that below some threshold the profits
available didn't support the growth of an "industry"
and, instead, favored fragmented markets and "one-shot"
products. What
happened in the late 1980's is, I believe, much the same as what
is happening now. Computer
science came into the picture.
"Disruptive" computer science. The
"Wave" that followed standalone PCs was networked
workstations and this "Wave" was led by Sun
Microsystems, Apollo Computer and others, once again a new group
of competitors. Then,
in sequence, what followed the "eight" year dominance of
workstations was, just as before, cheaper machines deploying the
breakthroughs of the previous cycle.
Yes, what Sun had proved to be practical on campuses of
thousands of networked machines was now "globalized"
with the construction of networks of millions of machines . . .
the Internet. Sun proved that "The Network is the
Computer." AOL,
Cisco and many other made it happen everywhere. (Much
of this was first published in 1989 and then made it into final
form in the infamous "rejected” HBR article titled
"Why IBM Failed," which was then published by
"Upside" in 1993 and can be read at http://www.upside.com/texis/mvm/story?id=34712c096.)
Once
again we are faced with the question, so what happens next? If
we presume for a moment that it can be demonstrated that the
workstation and Internet cycles also accounted for most of the
growth in their times and that these cycles also lasted roughly
eight years, then we have a series of five such cycles stretching
all the way back to the mid 1960's . . . indeed all the way back
to the dawn of the "magic" of Moore's Law. And,
if this is true, then it would appear that we are once again in a
similar circumstance to the one we faced in the mid-to-late 1980's
when standalone PCs had saturated their available market.
Now, networked PCs have clearly reached saturation in most
of the world. Looks
like a repeat to me! Is
it all over? Is the computer business basically mature and
therefore ready for consolidation . . . as claimed by some
regarding the planned HP and Compaq merger? Or,
as in the 1980's, are we once again going to drag
"disruptive" computer science out of the labs and begin
to build new platforms which fundamentally expand on the
capabilities of the multiple cycles of standalone and then
networked systems that are now behind us? I
am convinced that the computer business is *not* mature and that
"disruptive" computer science is already at work in our
industry. We
are on the brink of building computer/networking systems that will
enormously improve on the merger gains made recently in human
productivity. These machines will no longer require us to take care of
them, learn their complex ways and turn us into their mechanics.
Computers have given us a massive crisis of complexity that needs
to be solved . . . and we're about to solve it. IBM
calls this next step "Autonomic Computing."
MIT calls it "Human-Centered Computing."
They are both right. We
are now at the beginning of the first cycle of a new era of cycles
which will deliver computers that will truly revolutionize our
lives . . . or at least that's what I'm betting on. Mark
NOTE
FROM THE EDITOR To
summarize:
And
then I had a final thought, because I saw a reference to Gene’s
Law in a Dave Winer weblog and wrote Mark one more note: Applying
Gene’s Law to Likelihood of Computer Autonomy Mark: So
if amazing amounts of computer power continue to get smaller and
smaller (and therefore become part of ubiquitous, small devices)
-- see http://www.contextmag.com/magazine/setMagazineMain.asp
-- doesn't that mean these complex computers will by
definition require completely autonomic management? Amy
Amy: Yes
. . . the steady march of Moore's Law towards greater power in
smaller (and cheaper) packages does require some radically new
“architectures” so we can handle all this remarkable capacity. If
computing is ever to become ubiquitous . . . it will have to learn
how to take care of itself (i.e. what IBM calls "Autonomic
Computing") as well as learning how to “serve” humanity
in a much more seamless and productive fashion (i.e. what MIT
calls "Human-Centered Computing). This is where "disruptive" computer science comes
into the picture. What
we need (now that we have populated the world with complex
networked machines) is a new generation of computers that are
"Dog Stupid" . . . so that they can hang around in our
lives and reliably do important chores for us. Someone
once said, "If my dog was as smart as my computer, I could
take over the world" . . . well, the opposite is more likely
true. We need
computers that are as useful (and smart) as our dogs . . .and then
we could imagine living in a world that is truly an interesting
one. Mark
Comments or Questions: Send Email to
opinions@wohl.com
|
