Shane Greenstein: “How the Internet Became Commercial” | Talks at Google

Shane Greenstein: “How the Internet Became Commercial” | Talks at Google


SHANE GREENSTEIN: Thank
you for having me. Look, we’re small enough that
we can have a conversation. I do have a presentation,
but you can also interrupt me if you want. It’s really quite OK. I wrote this book, OK? Took a lot of years. And I hope to share a
little bit with you. I can’t share the whole thing
with you now, but perhaps get you to motivate
a little bit. And I really appreciate
you taking the time. So let me just say what
the book’s aspirations were and then give you
some flavor for it. The big question is why
did the commercialization of the internet have such
a large economic impact? That’s a big question. It’s the big question
in this area. What the book does, it
accomplishes three things. It puts everything
into one place I like to say there’s nothing
in this book that wasn’t already known by somebody. It just had never
been all together. Second, it focuses on innovation
and commercialisation. There’s lots of very good
writing on invention, a lot less on commercialization. And then third, it
offers a big picture, this phrase “innovation
from the edges,” which we can get to if you want. You might reasonably
ask, who cares? And Jonathan gave the
answer, because it changed everyday life. If you have children
and you talk to them, I don’t know if you’ve had
this experience of showing them “Leave It to Beaver on TV”
and then you say, look, look, there’s no internet there. And it really did change life. It’s changed it as we know it. It changed business
as we know it. It’s changed an enormous number
of industries, how they behave. It changed the identities of
the leaders in many industries. It’s the kind of
thing that it was responsible for a
boom in investment for about five years in the US. These are sort of things you
don’t see happening together very often in the 200-year
history of capitalism, if I can make an overstatement. It’s the sort of thing you
only see with electricity, automobiles, the big things. Indoor plumbing. I put a list there. Television, telephones. There aren’t many
examples like this, and so that, in and of
itself it’s of interest. But I thought it’s
also interesting just to understand what
happened and why, and why we had a big impact. You can also see some of
the symptoms in other things like all the
households that were online– you had half
the households online by the beginning of 2001,
which is extraordinarily fast, and the number of businesses
online 90% of US businesses were online by 2001, which
is also extraordinarily fast. The other place to
start is to start with misleading metaphors. The other reason to write this
book– many people, when they look at the internet,
want to look for an Edison or a
Manhattan Project, look for something that
the government did, as if the US government
orchestrated this, or somebody invented the whole thing. And that’s just wrong. That’s actually a very
misleading metaphor. There are some very,
very smart people who were involved in
inventing the internet. But if you want
to understand why it had the economic
impact it had, that’s not the place to start. That’s the other reason
to write a book like this. So I’m going to focus
on commercialization. This is the best picture
I could find of tubes. And commercialization,
in particular, is taking technology
and finding value in it. And that’s where the
focus of the book is. That’s typically not
something people write about, although you’re all
living it at the moment. The reason you want
to focus on innovation and commercialization,
it is a much better focus for understanding
creative destruction and the kinds of
processes we saw in the latter part of the 90s. And it’s also a much
better way to understand why some government policies
succeeded and others failed. So the big question that shows
up in the middle of the book is, how do major
technologies deploy? And what are the
processes that we observe and what are the
patterns we observe? And by the way, these are
very durable patterns, the kinds of commercial
patterns you find, again, over 100 years, the sort of
things you see in electricity, major agricultural inventions,
telephones, and so on. So if you start
from that question and you look at what happened
in the middle of the 1990s, you actually look at computing
as in sort of its 2.0 phase. We’re presently, by the way, in
about a 3.0 phase of big data. Actually, you’re at the
center of the present phase. I’d say historically, in about
the mid-90s, we were at 2.0. 1.0 was pushing out the
processors into the frontier and finding value in that. And the typical uses were
just the first times anybody’d ever done databases
for airlines, hotel reservations, simple logistics. And 2.0, a lot of people
knew that it was coming, was to do
internetworking, that is, to connect computers
over large geographies and do a lot of
automated processes. And a lot of prototypes had
been built. So the value was identified in advance. What wasn’t understood
was what form it would take in commercial markets. That was where a lot
of the mystery was. In this case, what happened
is a very good identification of processes you find in
lots of major technologies when they diffuse. And I would call them the
two conundrums in order to just identify for a general
audience what a lot of research sees over and over again. So the two conundrums
are, I call them the circular conundrum
and the adaptation conundrum. So when major
technologies diffuse, the big problem they
face is you have to coordinate multiple
suppliers all simultaneously around the same effort. And typically those suppliers
are competing with one another or they’re not even
talking to one another. And so getting
them to coordinate can be very challenging. The second conundrum,
adaptation conundrum, is when you have major
technologies diffusing, typically they
have to be adapted into a variety of circumstances
in order to be valuable, and that’s actually where most
of the investment takes place. And nobody wants to
make that investment unless they’re sure the darn
thing is going to pay off. And so then
everybody holds back. So you typically get
these two conundrums holding back major
technologies from deploying. And the question
that emerges if you look at any major
technology is, how do you resolve these two problems? So I’ll just do both
of these in this case. And I actually think it
helps people understand what we’re going through
today in a couple other major technologies. OK? That was the plan. So the circular
conundrum is often called the chicken-egg problem. If you’ve not heard
this, that’s a joke. I’m trying, all right? So the theory is, you have
to coordinate multiple firms. So how do you do that? The classic chicken-egg
problem has the characteristic that multiple firms need to get
their stuff to work together in order to create value. But all of them hold back, and
so they all sit on the fence, and so you get
long, long delays. Typically what you need
is either a focal point, a mandate, a platform
that coordinates, a standard that’s voluntary
that all can participate in. You need a mechanism
like that to generate an overcoming the
circular conundrum. I put a picture of
Google Maps up here because, in fact, that is
the function it serves. Just to give you an
example, this is still something we see today. So what happened
historically in ’95? Well, a metaphor that’s often
used is that what we got was a gold rush as the catalyst. That metaphor is
a bit misleading, but there’s a grain
of truth to it. And so I think walking
through the story actually really helps. Any of you ever been to the
place for the 1848 gold rush? Anybody? OK. So just to give you a
feel for what it is, that’s a picture
of Sutter’s Mill to give you an idea of what
gold rush economics is about. Gold rush economics is actually
an information problem. So what happened in 1848
is a good illustration. If you look at the
Sierra Mountains, they come up at an angle. They have the substrates
underneath the surface, and then the glacier cut
off a whole bunch of rock and exposed one seam of gold. Most of the seams of gold
that sit in the Sierras are not covered by rivers. But they are covered by
rivers in one place, which is the South Fork of the
American River, which Sutter built his mill on. And so what the
river does is, it took off a little bit of
gold and put it in the water. But nobody had ever gotten up
that high into the Sierras. Once Sutter put his mill
that high into the Sierras, the gold dust, which doesn’t
travel very far, became found. And what that did
is, very quickly information about that
discovery went out, and then we had a standard
economic behavior, which is it was thought
that getting to that gold quickly was going
to get high returns. So you have the
information goes out. Until then nothing is known. Then the information
spreads, and then there’s a behavioral reason for
everyone to rush in. That’s a gold rush. So the question is, did
we have a gold rush here? Was there something
unknown that then became known and then
generated a bunch of ideas that wasn’t known? And then what were the
behavioral reasons to rush in? That’s the sort of questions
you ask about the situation. So on the one hand,
if I was being very pedantic I’d tell you
this wasn’t a gold rush. There were plenty of firms who
were investing prior to 1995. And so you couldn’t
argue, necessarily, there was something that was unknown. On the other hand, if you’re
being a little more charitable and you look at
historically what happened, the coordinating mechanism
is pretty obvious. It was Netscape’s founding. And it wasn’t Netscape’s
founding per se. It was the demonstration
of a commercial browser as a commercial prototype. It was more than a prototype. It was actually for sale. And that was in
February of 1995. And essentially
what happened is, many participants in industry,
in the computing industry, at the same time saw the
same commercial prototype within several months
of one another. So you got the
equivalent of a gold rush because they were all informed
at roughly the same time, and then they all make
their investment decisions at roughly the same time. That’s actually
what happens here. We could talk about Microsoft
in a little bit if you want to. Microsoft has a slightly
different history at this moment. And Google might take a lesson
from that going forward. We’ll go to that
in a little bit. The other thing that happened
at the time, absolutely the one thing about this history that’s
essentially very interesting is that the governance
of the situation was unique relative to anything
that had come before it. Again, I think anybody
in computer science now takes this for granted, but
it was quite novel at the time. There was a governance structure
both in the IETF– sorry, in the Internet
Engineering Task Force– and in the World Wide
Web Consortium that had two characteristics that
had not previously ever shown up in a major platform. The two characteristics
were, anybody could come and gain information
about the technology. And second, there
were no what a lawyer would call restricted rights. There were no reach
through rights. The organizations giving
out the information did not tell you
what to do with it. So those two things. Anyone could come
in, and there was no constraint on what could
be done with the information. The internet platform
had no restraint on what was done
with the information. And that was just a
new characteristic. Any firm who had ever done
a major platform until that had always restricted
information flow in and out, and use afterwards. And what that did
is it generated what I think most
people in computing take for granted, which is
you get lots of specialists within platforms. So once you have a
well-designed modular platform, you can get a
specialist who takes for granted the
rest of the platform and then does a commercial
product that’s offered. In this era, this
is the first time we ever saw a major growth
across the entire economy in a whole bunch of
specialists taking advantage of the platform. I like to use this example from
Hotmail since it was so simple. Hotmail, the first
web-based email, and it had viral marketing. It had the little
footer that said, get your new email at Hotmail. And the people who
programmed this didn’t have to do an
enormous amount of work. They could take for granted
the rest of the network was going to work. TCP/IP was going to
work, the World Wide Web was going to work, and everybody
else was going to be just fine. And if they got a
wrong email address, it wasn’t going to
bring down the network. They could take for
granted they could do something very
narrow and specialized, and take for granted
everything else. And this is a great
example of what a modular platform with
unrestricted information can do, and did do in this case. So this was the other
source of the gold rush. So that’s the
circular conundrum. Ready for the
adaptation conundrum? The adaptation conundrum. I put up a picture
here of– I don’t know if anyone knows this one. It looks obscure, but it’s fun. Any Midwesterners in the room? This is corn. Particularly, this is
taken from hybrid corn. I’ll explain the
association in a minute. Again, here’s the theory. The theory is when you get a
new major technology deploying, it has to be adapted in
multiple circumstances. That’s something, again, we’ve
seen multiple times any time major technology shows up. What happened here, same thing. You would expect the internet
not to be used right away, but that it would need
a lot of co-investment in a bunch of locations
and a bunch of businesses to turn it into
something valuable. And then the open question
is, how does that co-invention organize itself? There’s no natural one
answer to that question in any major technology. And what happened
here– I sort of already gave you the answers–
we got a bunch of innovative specialists,
unlike what we had seen historically in some
major technology pushes, where one firm had
dominated the technology, say, for example as in
telephony, or in automobiles, the Ford Motor
Company dominated it. Here we had enormous numbers
of innovative specialists each doing their own thing. And that ended up
getting the investment in adaptation specialized
across multiple players. And it was also done very
quickly as a consequence. It meant all these people
and could act independently. So to give you an example,
why do specialists do adaptation really well? This is of my favorite
stories from this era. Did anybody ever
use this product? Any of you old enough to
remember this product? Internet in a box. It’s totally, you know, cute. So what they did–
this is a 1994 product. So what they did is they
took basically a browser. It was a Mosaic browser. It wasn’t even a
Netscape browser. They put it on a disk,
they stuck it in a box to make it look like
packaged software, because that was the
motif of the time. You would go to,
you remember Egghead or one of these retail
outlets, and then you would buy packaged software. And then people
went and bought it, and then they would put
the internet in their PC. I think it’s sort of cute. And you go, really? Come on. But this worked. This was extremely popular. And the point is that
innovative specialists all find different little
niche ways of adapting the technology to the
sub-market that they perceive. So they perceived
this sub-market for a bunch of new users that
nobody else had perceived. They got a huge
amount of attention. And so they were a hit, and
they eventually ended up selling to someone else. The key to an example like
this is the specialist is trying to learn something
you don’t otherwise see, you can’t learn in a lab. They’re typically trying
to understand something about the mix between
demand and costs that’s not otherwise learnable
through a simple experiment inside of a laboratory. Here’s another
example from the time. ISPs. Do you remember dial-up ISPs? Why did the US get the internet
earlier than any other country? This is the answer. Because we had the
fastest deployment of the dial-up network anywhere
in the developed world. And that arose, again, because
these dial-up companies were specialists in all
these different locations. And many of them had previously
been bulletin board systems. And so you found them
all over the place, and then becoming an
internet service provider was a very easy
thing for them to do. And so then this is my
favorite quote of the book. Let’s see if I can–
yeah, I can do this. “A good predictor of
not finding an ISP is the presence of a lot
of hybrid corn seed,” which comes from Zvi Grill, used
to be a professor at Harvard who studied hybrid corn. And this is his
1957 dissertation, one of the more famous studies
of major technology deployment. And he observed that ISPs
looked an awful lot like what he had studied years ago. You might ask a slightly
deeper question, which is where did all
these bulletin board systems come from? Where did all the
specialists come from? It’s one of the
characteristics of the US. It’s not unique to
the United States to have large communities of
what are known as wild ducks. Do you use that phrase
inside Google as well? This is a perhaps
archaic phrase now, a phrase about computing
about wild ducks, that people who look at
things in a different way or have a different
perception of what the innovative value is. The United States
is not entirely unique in having
communities of wild ducks, but had a fairly large
community even at this time. And there were a
bunch of regulations to protect them and mandate
that telephone companies should work with them. And then there was also the
First Amendment turned out to also support them, because
many bulletin board systems were doing some
unsavory activities. We’ll just leave it there. So I’ll finish up
here pretty quick. One other place that’s
interesting to find adaptation is in business use. There were two things going in
business that often are really under appreciated. One was on the browser side,
which looked a lot like home. The other was on enhancement
of business services to support electronic commerce. That was really expensive,
that kind of investment. As it turned out,
after the fact– we shouldn’t be
surprised by this, but many people at the
time were surprised after the fact– there was
a large incentive by users to retrofit existing processes
with something that saved the capital they already had. Users did not want to go
into a green field situation, typically. They typically wanted
to preserve a lot of the things they already had. And as a consequence,
many of the dot coms who came into these
businesses failed precisely because they didn’t respect
their end customer wanting to preserve what
they were doing. And if you look in
retrospect, the reason IBM succeeded in this
era was precisely because they
respected the existing processes of their client base. That’s just one
of the big lessons you have to walk away
with from looking at this era, for
what it’s worth. We could talk about
the boom if you like. There was a boom. The boom had a gold rush. It was caused by the gold rush. There was a second
thing going on, and that was a
network effect where investment by firms in processes
made browsing more valuable. That made more firms go online
with electronic commerce, which then induced more browsing
adoption, and so on, and more investment in the network. And so you had these
interplays which made it more valuable for each
player to do more investment. That was the second thing going
on, and that caused the boom in the ’90-2000 era. And then, as an economy, we
overshot, without question. So that’s, in a nutshell,
chapters nine, 10, 11 and 12. Just explained how that
overshooting happened. So I’m doing the whole
thing quickly just because it’s more fun. Let me end with two things. Renewal. So after the overshooting
there were two forms of renewal that I talk about in the book. One of them is Wi-Fi. And where did Wi-Fi come from? Wi-Fi is a wonderful example
because the spectrum was allocated initially for things
that the engineers in the FCC regarded as garbage, and many
engineers regarded as garbage. So baby monitors,
mobile handsets. You remember these kinds
of things inside your home? Garage door openers. And the trick, the
thing that happened was, the unlicensed spectrum
allowed equipment firms to embed the use
of that spectrum inside of their wireless
antennas and their wireless servers. And users found
that very valuable. And so the spectrum
ended up migrating from garage door openers. You still find it in
some garage door openers, but it ended up migrating
mostly into the area of Wi-Fi. And so the unlicensed
spectrum ended up being the vehicle
for movement of value from a low value use
to a high value use. And that’s a pretty
interesting piece of renewal. And then I had to do this
example in this room, just because it’s well-known. The book also has a chapter
on internet advertising and Google’s approach to
renewing that market, which was on its way down at the
time that Google starts to figure out how to do an ad
auction for keywords, which renews the market for
advertising at the time. And I suspect this
story is known in here, so I won’t go over. It’s just I had to bring it
up because it’s in the book. Want to do big lessons,
or are we good? Big lessons? OK, big lessons. That’s my favorite big picture. Sorry. That’s also a bad joke. OK, what movie? Yeah. Ferris Bueller. So the big picture. Innovation from the edges
is the major framing. That’s about outsiders bringing
new perception and new assets to bear on an opportunity that
insiders were not investing in. The big lesson from
this experience is the way outsiders
explored, particularly outsiders who were specialists,
and innovative specialists explored sub-pieces. And that unlocked the circular
conundrum and the adaptation conundrum. That’s the big insight
of this experience. The book spends a lot of time
filling in the gaps about why that happened and why. It has a lot of consequence
for thinking about, say, for example, a company
like this who supports a lot of innovative specialists
as a major platform, and how would you think about
taking activity to support? And then start new
platforms to support new innovative specialists. So in this instance,
what we observed was this interplay between
all the specialists who raised the value
of the investment by all of the others. And the other big lesson,
in comparison, why did you have this done in this way? Why did the network
grow this way, and why didn’t it grow
from a telephone company? There were lots of
opportunities for publicly supported telephone companies
to grow a data network. And why didn’t it
happen that way? And the answer is that
innovative specialists perceived things that
otherwise would not rise to the top of priorities
inside of a large firm. You get multiple points
of view supported in this kind of market structure
that you wouldn’t otherwise get in a single firm. And as a consequence,
you get more variety than you would otherwise
get in a single firm. But if you were going to read
this book with the present in mind, I would
start by saying you would expect to see these two
conundrums arising repeatedly. They still do. For example, if you look
at big data 3.0– as I say, it looks to me like we’re
in 3.0 of big data– we’ve been in big data for a
long time as an industry. But the present era
of big data, you would expect to see circular
conundrum issues showing up, which you do. And you see major firms
trying to take positions as coordinators. I think you’re in that business. So is Amazon. So is Microsoft. You see institutions that help
coordinate large data as well. Again, we tend to have a lot
of open systems doing that. You see government
mandates, again, trying to take positions to
unlock circular conundrums, particularly in the
Health Care Act. There was actually lots of
subsidies in the Health Care Act to get hospitals
all to do computing inside their organizations
so that then they would all use the same field
so they could exchange data. That’s a very standard way to
overcome a circular conundrum. Then on adaptation, you would
also expect to see that. You see that again. I mean, I could run
through it again, but you see that in the
present environment. My own forecast on say, for
example, the big data changes today, is that adaptation
is actually much easier than we saw 20
years ago, partially because you don’t really have
to involve households very much. Most of investments being
done in the infrastructure are behind the scenes, and it’s
being done by major businesses. And so largely those investments
are done for private reasons and don’t need to be
coordinated very heavily with a lot of other firms. There’s some standardisation
issues, of course, and some big government privacy
issues that are difficult. It’s a challenging problem,
but it’s easier than the one we just observed in this book. Thanks. [APPLAUSE] I’m happy to have
a conversation. It’s small enough in here. AUDIENCE: Could you
give a specific example of a circular conundrum
that you [INAUDIBLE]? SHANE GREENSTEIN: Oh, certainly
on the smartphone market. I mean, yeah, though we’re
over that one at this point. In smart phones 10 years ago,
we were still– think about it, before the iPhone
we were still facing the same old circular conundrum. Nokia and Microsoft had both
done a lot of investments to try to overcome that. And then the iPhone ended
up being the catalyst to start things, and
Android was close enough behind, and with a
governance structure that was very friendly to
programmers and app developers that it also managed
to develop a network. That’s what you have in mind? AUDIENCE: Yeah,
well, how about one that hasn’t been resolved yet? SHANE GREENSTEIN: Oh,
that hasn’t been resolved. Certainly internet of
things has got that in it, because of a bunch of the
standards about security, for example, and how things
should work with each other when they come across
from different firms. AUDIENCE: Do you see the
emerging television standards, now that everything is going
on the internet, essentially? There are already beginning
to be interoperability issues between smart TVs
and Roku and so forth. And in fact, I was just on the
other side of the building, and people were talking about
the possibility of testing applications on things that have
new models out every season. SHANE GREENSTEIN: Yeah, just
in the streaming video side, even in the consumer
side, there’s still some serious issues. You know, there’s actually
some public policy there also in getting the coordination
between the– I almost hesitate to get into
the weeds on this– on edge providers, if you will,
and broadband providers, and what’s appropriate behavior
at handoff points and gateways between network providers. If you look, for example,
at the issues that arose when Netflix was not
working too well in January of 2014, to use a
very simple example, that was from not
having an understanding about how to coordinate data
handoff between, at that point, either a backbone provider
who was carrying a lot of data and passing it into
a broadband network, or from a CDN into
a broadband network. And there was a disagreement
about the proper way to do that, or what both
parties wanted to do. That’s a pretty nasty
circular conundrum effect, if one keeps arising, especially
if everything goes online, as we’re all forecasting. That’s another one. Is that helpful? AUDIENCE: Yeah. Well, I mean, are
there any broad trends that you can use to predict
how, say, internet of things– SHANE GREENSTEIN: I mean, cars,
which your firm is involved in, is a pretty good one as well. We can all see that one coming. There’s more than just– I mean,
the cars already work, right? Yeah. They already work. So the prototypes–
we’re in a situation now where the
prototypes all work. We can all forecast
there’s going to be a scale decline in cost pretty soon. And there’s an enormous
amount of coordination that needs to take place
between road building, insurance companies, legal standards
for accidents, and liability. AUDIENCE: But in
this case, it doesn’t seem like firms are
waiting to invest, right? There’s already lots of
companies investing in that. SHANE GREENSTEIN: Yes. There is already a
lot of investment. AUDIENCE: Is it that there isn’t
a clear direction, so there’s– SHANE GREENSTEIN: Yeah. So the danger at the moment
in that area is we’re going to get
Balkanized standards. So we’ll get two or three
different ways of doing things. In fact, it almost seems
inevitable Europe will go one way and the US will go another. That just almost seems
inevitable on this one. Even worse is California
goes one way and the rest of the country goes another. That seems possible also,
particular with cars, because, California has always
gone its own way on, say, for example pollution controls. AUDIENCE: So in this
case, so for cars, government
regulations are what’s going to ultimately determine– SHANE GREENSTEIN: It’s going
to be a major determinant. AUDIENCE: Is that
always the case? Like for Internet of Things,
is there another way it go? Can you predict how the
net will be resolved? SHANE GREENSTEIN: No,
it’s hard to predict. I never like to predict. I would be a little cautious. In this example, 20 years
ago, government regulation had two roles, and so
we should be careful. Sometimes it’s very
interventionist because competition
policy intervenes very directly when you
have a monopoly provision. That’s still going to
happen in this country. There’s a longstanding
dislike for monopoly provision in the United States. That’s not going to go away. Monopoly provision is
going to happen somewhere in transportation because there
always are, it always does. And you can forecast there will
be government intervention very directly in the places where
monopoly provision of transport services allows somebody to
jack up prices or dictate terms, and government regulation
will intervene. Just know that, because that
there’s a long history of that here. Having said that, one of
the interesting things about watching the
internet experience is, government policy was
not orchestrating. It was often stepping back. It was very deliberately
stepping back and letting firms invest as discretion dictated. And you see a lot of that in
the history here of particularly federal forbearance is
the word the lawyers use, a deliberate stepping back
from making decisions. So it would enable
private industry to choose what it wanted
to do without orchestrating the whole thing. And it was very
selective intervention in things like establishing
the Internet Engineering Task Force, or in privatizing
the backbone. But it was very selective. I would say it was always wise. I mean, the domain name system
wasn’t done particularly well, for example. So there’s something
equivalent’s got to happen in cars. There’s got to be a set
of standards, for example, on how information is
going to be interchanged between the various
parties, particularly between the people who watch
roads and individual cars. Some of the things–
I mean, you guys would actually know
better, I suspect. But if you’ve watched some
of the prototypes I’ve seen, you get that you get
real time communication between the vehicle and
some other public source of information. And then that
requires something. AUDIENCE: Your colleague
Professor Christianson argues [INAUDIBLE] and
how companies basically get stuck at [INAUDIBLE]. You were saying
that specialists are need to resolve that dilemma. SHANE GREENSTEIN: Can help
resolve that dilemma, yes. AUDIENCE: How do you see
companies or leading businesses such as Google avoid that
dilemma of having specialists inside? SHANE GREENSTEIN: I was
ready for this question. I brought a slide. So can I do the history first? If I had to take a history. OK, I have an Al Gore slide,
too, in case you’re curious. OK. So if I was going to take–
I know people at Google might resist being compared
to Microsoft 20 years ago. But work with it for a minute. Microsoft was
approximately 20 years old when these events occurred. Not saying– Google’s
not exactly 20 years old, but it’s pretty close. There was some very
strong personalities at the heart of that
firm at the time. I’m not saying that’s
particularly true here. But there’s a very large firm. What happened to Microsoft
in this situation, the details are well known. Gates had lots of
personal authority. And he had misunderstood the
potential for the internet. And there was a
skunkworks inside of the company that had
done a lot of advanced work. And then I put a
picture of Ben Slivka up here because he’s actually
the guy who wrote the memo that grabbed Gates’ attention. And they grabbed
his attention late. So that that’s a
historical fact. He came to understand what
was about to happen later than many other in industry. If you look at that example
and step back from it and say, what’s the lesson, which is
the essence of your question, I would sort of be both
empathetic and critical at the same time. If you look at
Microsoft at the time, they were very good at what you
would think of as a big push, deploying products that took
several thousand people, organized over multiple years,
towards a very big goal. They were extraordinarily
good at that. And you know,
specialists actually can’t do that very
well typically. So it was a valuable
thing for them to do. The best example we have
at the time was Windows 95 actually fits this. What were they
very poor at doing? And I would guess the same
would be true here as well. There are sort of two
things they were not very good at doing. And it’s just inherently true. Every large firm
is poor doing this. They weren’t very
good at planning. They were extraordinarily
good at planning relative to their rivals, but
they still weren’t very good at seeing the
future, because everybody’s poor at doing that. And second, they resisted
cannibalizing their own product lines. And lots of reasons why. We could go into those reasons,
but lots of reasons why. And it led them in the
direction that you’re hinting, to resist understanding
businesses that were inconsistent with
their present business. That was the actual
thing that happened here. If you look at the mistake
Gates made at the time, it was, he didn’t want to
cannibalize Windows 95. And he understood its value
in a very particular way and resisted another
understanding that was inconsistent with that. And the result of that
was it slowed them down a tremendous amount. And even for a long time after
Gates’ change of direction, he resisted. He actually did
resist investments to take advantage of what the
commercial internet would have allowed his firm to do, because
he just resisted cannibalizing his own investments. That, I think, is
the big danger. It’s a big danger
in a firm like this. There’s just natural reasons
why existing firms hang on to existing product
markets, existing revenue sources, and
existing perceptions about where the source
of value comes from. Yeah, that’s the issue. And then the hard
part– the hard part is getting the timing right. Gates was late. That was a preventable
error in his case. I think, in practice, I actually
want to be more forgiving. It’s actually very hard to
get timing right in general. But that’s the
lesson of this case. You want the Al Gore slide? OK. Sure. Everybody always asks. Nobody’s asked yet. But I’m always prepared for this
because everyone always asks. What did Al Gore invent
and when did he invent it? The book– actually,
I felt an obligation to figure out what
actually happened here. This is the actual quote
that started it all. This is a Wolf Blitzer
interview in 1999. You can read it
as well as I can, but it’s “I will be offering my
vision when my campaign begins, and it will be
comprehensive and sweeping. And I hope that it will
be compelling enough to draw people toward it. I feel that it will be, but it
will emerge from my dialogue with the American people. I’ve traveled to every
part of this country during the last six years. During my service in the
United States Congress I took the initiative in
creating the internet. I took the initiative
in moving forward a whole range of initiatives,”
et cetera, et cetera. That’s what did it. So after that
interview, the next day, there was a bunch of opposition
research, most famous of which came from Trent Lott. “If Al Gore invented
the internet, I invented the paper clip.” Dan Quayle was my
favorite quote from here. “If Al Gore invented
the internet, then I invented
the spell checker.” And that’s how the meme started. And then this meme got started
on late night television. Top 10 things Al Gore invented. And it just gained
its own momentum. And the ridicule is
quite funny, because it’s obvious no one individual could
have invented the internet, nor could one policymaker
orchestrate the entire thing. That’s sort of
inherently ridiculous. And that’s why it was funny. It was also made for good
political campaign, opposition campaign. But that’s its source. And the interesting
thing, actually, watching this in
retrospect– try to explain this to your kids,
to someone who wasn’t there. How did this survive
for so long as a meme? And that’s the part that’s
actually hard to explain. And it was just
very good politics. And this politician lost
control of the conversation. That’s where it comes from. It’s a great joke, though. AUDIENCE: But he did push
through legislation that– SHANE GREENSTEIN: Yes. So his actual
legislative history, if you want that– his
actual legislative history is two things. It’s two pieces of legislation. It’s primarily, though, with
funding the National Science Foundation network, particularly
the latter bill that upgraded the backbone for
the internet and funded supercomputer centers,
which were attached to it. And then one of those
supercomputer centers was the source of
Mosaic, the browser that generated the
prototype browser that was the catalyst for
the commercial browser. And that backbone was the
backbone that was privatized, that generated
the other movement towards commercial internet. So yeah, he deserves
credit for that. That’s a real thing. How’s that? Cool. Well, thank you very much. [APPLAUSE]

Leave a Reply

Your email address will not be published. Required fields are marked *