together we can change ourself

together we can change ourself

Keynote talk @ Stanford conference — Vinod Khosla

leave a comment »

First let me start by saying that I don’t know the Indian environment as
well as most of you do. So if some of my ideas seem naïve, they probably
are. There are some fundamental principles that I would like to talk about
that are quite different from what has been done in the past. But I do
sincerely apologize if some of the ideas in fact are naïve. I put this
presentation together this week specifically for this conference and so it
probably hasn’t gone through the iterative cycle of finding and plugging
holes. Having apologized let me suggest that many of the traditional
models of why we need telecommunications may in fact not be correct.
Let me first start very quickly by talking about why we need
telecommunications and what implications it has for a country like India.
As to be clear on what we are building a telecommunications system for. If
you look at what’s happening in the US, there are a number of things
happening that are non-traditional. 40% of GDP growth coming out of very
small industry sector. Morgan Stanley is predicting that a 100% of GDP
growth will come out of this sector within a few years. That’s a pretty
amazing fact. I have had many a conversation around how deflationary a
force technology is. None of the econometric models that relate growth
rates to unemployment and interest rates are not working. In fact almost
all the models have been dead wrong because the historical relationships no
longer hold. And I can’t tell you what the right model is, but what’s very
very clear is that the current model is wrong and it’s wrong for the
following reasons. Every model I know includes the price of oil. None of
the models include the price of microprocessors or bandwidth. When
microprocessors and bandwidth are driving the economy as much as oil, then
the model is clearly wrong. The other thing that we have done in the US
that’s surprising is harness entrepreneurial energy. It accounts for 40%
of GPD growth. Let me just speak for Kleiner-Perkins for a minute. There is
probably no more than a dozen people who in the last 20-25 years have
helped start at least 10 of the Forbes 500 companies. So that large a
dislocation, not counting all the companies that don’t quite make it to the
Forbes 500 list, can come out of such a small group. It comes from enabling
entrepreneurial energy and I will talk more about it.

This is really testimony to the power of ideas and I am a real believer in
the power of ideas. It is testimony to the power of entrepreneurship and
probably equally important to the power of technology to impact on our
lives. I will come back to this in a little bit. I think the transformation
that we are going to see in India has not started and I have not seen one
economic paper on why or how we might handle this transformation. At the
turn of the century, about 100 years ago, most of the employment in the US
was in agriculture- 50% or so. Its true of India today. The fact is less
than 2-3% of the population is required in agriculture with today’s
technology to support the remaining 97% of the population. So we should
plan as a country within a relatively short time to have most of that
rural population to be doing something else. Not catching the problems we
see today, but how to smoothen the transition to a state where only 3% of
the Indian population is working in agriculture. Because we know that the
technology is there to have that percentage produce enough agricultural
products to support the rest of the country. That’s a fundamental fact. By
the way when land was the scarce commodity, land ownership was the source
of wealth and power. About 30 years ago in the 70s we saw the industrial
economy, the manufacturing sector (Japan is a classic example country that
leveraged that part of the economy) accounted for the bulk of employment
more than 50% in the US. Oil and factory ownership was the source of wealth
and power and capital was the scarce commodity. Japan in fact had very very
low interest rates compared to the rest of the world and used that,
together with a high savings rate, to leverage itself into winning the
manufacturing race and transformed their country (which was pretty close to
a broken country after the second world war) after the war into a major
power house. Let me suggest that in the US less than 10% of the population
will be in the industrial economy 25 years from now. The US has to go
through this transition. Massive dislocation of skills – we saw that in
the industrial sector of New Hampshire here a while ago in industries like
textiles. Let me suggest that the new scarce commodity will be human
capital. I will come back to this.
There is a number of requirements for success in this knowledge economy and
the reason I am harping on it is because I don’t believe that a country
like India can transform itself by chasing old goals. Taiwan and China are
going after factories, production of plastics and little things. Let me
suggest the way to transform the Indian economy is to take the new growth
segment of the international economy – the knowledge economy where India is
relatively well positioned and deploy its human capital against this scarce
resource. And let me suggest later that the way we design our
telecommunications systems is to enable this redeployment. I won’t go
through this in detail. The one thing I would say about this is education,
life long learning, entrepreneurship, those are key factors. Enabling this
is the roles governments will have to play. I will talk about the economic
models of how you make the transformation and how you make it socially
acceptable and politically acceptable later.
But what do we really mean by this? Let me pick a couple of examples. Let
me suggest that what freeways were in the US in the 30s and 40s, what oil
was in the 60s and 70s bandwidth is to the new economy for next 20 to 30
years. The principle resource we need more than capital is human capital
and that in fact it is in abundant supply. All we need to do is learn to
harness it. It is almost like solving the power problem by using solar
energy. We haven’t figured out how to do it. But when we do, we have solved
many fundamental energy problems.

What are the components required to achieve this transformation? I believe
as I said that bandwidth and human capital are the raw material (there is
plenty of capital around and capital is no longer the scarce resource) and
importantly culture. To me silicon valley has now become a culture, a state
of mind. It is not a place anymore. I was in England recently, in the south
of England about 3 hours outside of London. In the space of one morning I
heard three times about how they were like silicon valley. It is truly a
cultural thing and frankly if we have all these ingredients lined up we
have all the components for massive transformation. And we should design
our telecommunications system to enable this to happen.

I had recently an engineer leave one of my companies. He was collecting
stock options worth a few million dollars a month. He was a critical
engineer, and I said, “You are leaving a lot of money on the table” and he
said “I can make more somewhere else”
. This is why ideas and intellectual
capital have become so important. It is about intellectual capital and not
about financial capital.
More and more this economy is being driven this
way. For those of you who wonder what most of us do, I spend probably
30-40% of my time recruiting because human capital is the hardest resource.
And of course human capital is more available in a place like India, if we
can enable it, empower it. The other great thing about ideas as a source of
capital- remember I talked about land ownership being a source of power and
wealth in the old days and then it was oil or ownership of industrial
factories later – ideas are egalitarian. They cant be inherited. That’s
what is great about them. They come from anybody and any place. They have
to be powered by entrepreneurial energy. But they are not class based. They
don’t depend on who you are or which family you come from or how much
wealth you have and they are becoming the most powerful asset in the new
economy. I am going to suggest that we restructure the way that society is
organized and suggest a couple practical more evolutionary ways to do it
rather than radical ways where you dictate everything. Technology has given
us increasingly more powerful ways, from the first stone tools we fashioned
as humans, we have had increasingly more powerful technologies and we have
used it in a number of very productive ways. But most of our society is
organized around very old communication technologies. So what are the most
important communication technologies of the past? Think about how this
country, India or how any other country is organized. It is organized
around 300-400 year old communication technologies. It happens to be
rivers, mountain passes and ports. Cities are raised that way , trading
patterns are set up along these lines, relationships are built along those
lines. Let me suggest that the internet is the first mass 2-way
communication technology. And probably the single most important technology
since the printing press. The printing press was important because it was
the first accumulation of human knowledge that was made possible. You could
handwrite books before that but you could not disperse knowledge on a
practical scale. And you couldn’t build on other people’s written
knowledge. That’s why the printing press was important. Let me suggest that
the internet is the first 2-way communication medium that is starting to
get rid of many of the physical constraints – rivers, ports and mountain
passes around which our society is organized. This will happen anyway but
we can facilitate it. This technology is at least as important as the
invention of cars and TVs and airplanes if not more important.
We can look to technology and say that this is a real reason of hope. I
spent 3 years in India, I went there in 1993, committed to spending half my
time here and half there, every 6 weeks, I must have been the biggest
commuter in the world, and I thought that I would look at what I could do
to help there. Within a couple of years I was very depressed because it
seemed hopeless. The scale of the problems were too large to solve in any
practical, rational way. Since then I have been thinking about the problem
and I realize that (Paul Romer is not here – he sort of coined the term
increasing return economics), we have to set in place a mechanism for
increasing return economics. If we are going to solve these problems in any
sort of reasonable time frame that is the only way. The returns will have
to grow exponentially. But we look to technology and its impact – in
agriculture, in health care, in communications, in transportation in gives
me hope that we can replicate these increasing return economics.
So let me now talk specifically about telecommunications and let me talk
about first the rural problem and then the urban problem. Let me try and
address them both and suggest a couple of ideas none of which are fully
refined and probably have lots of holes but are worth thinking about. First
one can look at it linearly- do we have enough resources, how many
billions. You know a trillion dollars which we could never raise is only a
thousand dollars per person in this country. Clearly linear extrapolations
don’t work. Can we discover a new oil? Like they did in Saudi Arabia. I was
looking for all those solutions that might work. What we need is a
miracle. I will conclude by telling you that a miracle is possible. I have
looked at the model of foundations that give money away, but there is not
enough money you can give away. We need trillions. So once we conclude that
we have a big problem and there is no conceivable way to get enough
resources, the next thing is to say what is the best we can do? What’s the
pragmatic approach? One of the things I hear about is the digital divide.
It’s a good conversation to have in the US. It is not an appropriate
conversation to have in India. Let me suggest that what we don’t want to do
is to give everybody access to telecommunications. I know it sounds like a
radical thought. Let me suggest that we want to do is take our resources,
whatever little we have, and put them to the highest possible economic use
for those resources. And if you believe in the entrepreneurial model like I
do, I believe that 5% of the people empowered by the right tools can pull
the remaining 95% of the country along in a very radical way. And in fact a
few years ago I looked at this country and said, how many people accounted
for the bulk of the social and economic changes in the last decades. There
was no more than a few thousand people who drove that change. You can look
to a few people, you can look at people in academia, people in business or
political activist. No more than a few thousand people accounted for the
bulk of the change. Not for doing it but for driving it in this country. If
we can take our limited resources and give them to 5% of the people and I
will come back to how we pick which 5% percent as that’s the politically
important question, and let them put it to the highest use so that we
multiply these resources. When we start with a dollar, next year we don’t
have 50 cents to give away, but we have 5 dollars to give away. Because
entrepreneurial energy and a business case is being applied. Let me do a
simple calculation for you. I originally did this calculation in the US
context. For 3% real growth accumulated over a 100 years results in a
multiplication factor of over 20. So if you assume the US per capita income
is, pick a number, say 35,000 dollars and we accumulate 3% growth for a 100
years, the per capita income in the US at the turn of the century will be
over 700 thousand dollars. That’s a humungous number. Why? We are applying
the exponential equation at 3%. At 3%! We can clearly achieve much larger
figures in the context of India if we use these resources effectively.
So which is the 5%? If you believe in the entrepreneurial model the 5%
isn’t the richest 5%, or the most privileged 5% or the people who can
afford it the most, it’s the people who can use it best. I will use this to
suggest a new model. So let me go on to rural access. There’s 500,000
villages in India. Let me suggest that trying to empower them with
telecommunication is a bad idea. Its uneconomic. I would like to start by
explaining why it’s a bad idea by examining what has been happening and
what is likely to happen. Every nation, in India’s situation- India,
Mexico, Argentina, everywhere in the world, they are trying all these
programs. What we are getting is very few resources in the rural areas
despite years of trying and good intent, are sprawling cities, climbing to
10 or 20 million people. A few major cities – Buenos Aries, Mexico City,
Bombay- and the bulk of the population is living in these sprawling slums.
And the villages lack power, communications, infrastructure, education ,
health care. All these problems need to be addressed and need to be
addressed simultaneously. You know if there is no conceivable way to do all
these things in every village, what can we do? It is fundamentally
economically unattainable to try and do what we have all been talking
about. Talking about rural lines to a village of 100 people or 100
families. It is not reasonable to even shoot for those goals. Here is a
suggested model:
I roughly took a map of India and said what if we drew 5000 circles, 40 km
in radius. You can call them blocks, you can call the them districts. I
didn’t want to use any of the traditional terms for a set of reasons and I
call them circles. Each circle covers a 100 villages, about 5000 sq kms
each. I estimated there were a 100 villages in each, on average 25000
families- 100,000 people in each circle. Now that’s a viable scale at which
to build many of these facilities. Out of a circle of 100,000 people I can
see a few thousand people effectively using all these technologies we are
talking about. 1,2,3,4,5% of the population. By the way its important to
say that there is viable real infrastructure already in place to support
this circle idea. That’s the other really expensive infrastructure to
build. And we combine that and multiply it with the telecommunications
infrastructure. We try and build 5000 important centers and not 500,000
villages. We build it in a way that anybody, any of the 100,000 people in
the centers have access, but only if they are motivated, only if they are
driven, only if they are willing to work hard and if they have a good use
for this that has a good ROI. Now plenty of details have to be worked out,
there are plenty of economists here who can work this highest use model. We
define universal access in a different way. Not that everybody gets a
telephone at home. But they will have access, not only, by telephone but
also internet access and health care and education if they have the
motivation, the drive the hard work because that’s the energy we need to
leverage in a country like India. It is socially viable. In the current
model when somebody moves from his village to Bombay, a three-day train
journey away, away from their family, their support structure, it results
in socially undesireable behavior- like drinking, crime, not caring- the
things that are the cause of problems. It results in slums that we have all
seen in Bombay, Delhi, Mexico City– any big city. This new format is
socially viable because you keep the traditional infrastructure intact.
People may move 40 km at the most, on the average 10-20 kms, a bicycle
commute away. You don’t need much more. This won’t cause radical shifts. I
am proposing what we do is engineer this and engineer the design of a
country like India by setting up these circles, setting up the
communications right by the rail and you enable this migration to happen
locally, gradually and slowly. So 5000 circles support a 100,000 people per
circle. 25,000 families or 100 villages however you want to look at it. But
what it does is it does it over the course of 10 years. By the way 5000
circles and 100,000 people is 500 million people and I will come back to
the remaining “urban” 500 million. I did a nominal budget to get setup
these 5000 centers too. The social process of change by putting the honey
where you want the bees to go is a gradual slow process. In fact I would
suggest that it is much less dislocation than what is happening today where
people move from their village to cities many thousands of kms away and
have no reasonable way to get back or communicate with their families. I
would allow this gradual process to happen and instead of having a country
25 years from now which has 10 cities with 30 million people in each city,
we will end up with 5000 small cities, a reasonable distribution of
population, socially more productive, socially more supportive.
Sociologists can talk about how much more attractive this model can be. I
think I have talked about this issue of teledensity. Its available but not
granted. It is shared because we don’t have the resources to give a
telephone to every person in India. It’s economically productive, internal
rate of return oriented and allocated to the highest economic use.
If we said that we wanted to get an earth station or fiber in every one of
these circles right by the railway center and invest another million
dollars, a million dollars per center- that’s only 5 billion dollars- to
put a million dollars of infrastructure around every 100 Indian villages,
and empowering all 100,000 people to do their best and get paid for their
entrepreneurial energy. They will have through these circles, work centers,
communication centers, access to worldwide markets. I read this great
article in January in India Today, about how they were trading milk over
the internet. And you enable not only that but reach into the whole world.
A person I know runs a charitable organization that makes these great hand
crafted wooden toys in India and they can’t monetize them. They can produce
as many and I know that you can sell as many of these toys as possible at
$10/a pop here in the US. But there is no market place mechanism to match
the demand and supply. I can’t tell you which uses and we shouldn’t plan on
uses and I find studies on the use of internet almost silly because they
never match reality. People discover the uses, the discovery process is
what is important. You ask anybody and every pundit in America was
forecasting interactive TV as the use for communications 6 or 7 years ago.
By the way when you have economic centers like this, you can have health
care, you can have online education resources, and how do you get 5 billion
dollars? If the communication market in India is 8 billion dollars a year
or 32000 crores is the number I heard recently, a 12% tax gets you there in
few years. It’s a percentage tax and not a rupee tax because you want
telecommunications entrepreneurs to keep building the telecommunications
business and growing the top line the RS.32000 crores as rapidly as
possible. I will come back to the economic model in a better way.
It is not hard to see how you transform the country in less than 10 years.
Completely. So now let me switch to urban India. Again I want to state that
these are ideas I have been thinking about for sometime now, but put down
to paper the for the first time this week. So there’s lots of holes and I
hope all of you will give me feed back on not only the holes but also
suggestion on how to fix them. Let me suggest that the US model for
competition is not the right model for a country like India. If you look at
telecom infrastructure in this country and many countries there is 2
distinct parts of it. There is the very static infrastructure and then
there is the very dynamic piece. When you have the right economic model and
not enough resources you share all the static pieces. Why are rail, power
and gas are all building their own conduits, duplicating resources, I
don’t understand. It is a real waste of money that we don’t have. Why not
have one joint effort to put in 12 or 16 conduits, however many we want to
pick everywhere in urban India. While we are doing it we can put in
underground electric, underground gas, all those other things because I am
pretty sure thirty years from now the nature of the conduit would not
change. In the process of doing this let us not do what we have done in the
DoT before which is to decide what fiber to put in. Let that be the dynamic
piece. Let’s not decide what equipment to put in at the ends of these
conduits. Same thing for wireless towers. Why shouldn’t all wireless towers
be shared? Because that doesn’t change in 10 or 20 years. The technology
that goes on top of the towers- absolutely. Let people innovate, change be
dynamic about. Let people have 20 year leases on these towers and conduits
so that they can keep changing and competing and beating each other’s
brains out, so that the effective cost of telecommunications goes down
rapidly. We know that the cost per unit bandwidth per bit is going down
very very dramatically. We have an artificial situation in the US in that
nobody expected the demand to be so high and its going down at only 30-40%
a year. But the technology is there to support declines of 60-70% a year
and once production catches up with demand we will see much more radical
declines. I can’t imagine why anybody would ever want to bill a voice call
by the minute. Today it costs far more to bill a call than to provide a
call. Billing for a 5 minute phone call costs as much as providing for a 50
minute phone call. That’s broken and something is wrong. So we allow for
economic efficiency on the static infrastructure, we allow for competition,
experimentation, success and failure in the dynamic pieces and we enable
Couple of other thoughts. Nothing should be a fixed number. If we decide we
want to combine the rail and gas and power infrastructures and put one set
of conduits, that’s great. Lets charge a percentage fee, lets charge a
percentage fee on the wireless spectrum. You know people always argue about
rates, but whether its 5% or 15% is not relevant. If we want to be more
aggressive about social redistribution, lets charge 15, if we want to be
less aggressive, lets charge 5. That’s a policy question. But as long as
there is a percentage, it lets the entrepreneurial energy apply to growing
the entire market segment. The total pie grows and the faster it grows the
faster your percentage grows. So every thing should be structured to
encourage competition. Encourage optical, wireless, all these technologies.
A couple of other thoughts on technology. Let me suggest that it is brain
dead to build a PSTN infrastructure on India. It is the wrong thing to do.
We know that the world is going VoIP. It is an order of magnitude cheaper.
Why not go that way? Lets leave things as the are. Lets not enhance that
infrastructure. Lets not make any more investment in the current
infrastructure and India will have a huge advantage. We don’t have
expectations of quality like we do in the US. We don’t need 6 nines
availability. We got 30% availability already and that’s all we got. This
infrastructure is ready to support reasonable availability and we know that
people will buy that even in the US. While AT&T tries to provide 99.9999%
in reliability 5 times AT&T’s capitalization has been built on the wireless
infrastructure which is far from 99%. And we know that people use it and
it’s the fastest growing segment in this country. VoIP can do much better
than wireless. Why don’t we wholesale switch to that technology. Why even
have a regular phone? Lets not do media gateways like we have to do in the
US because most of the phones are standard analog phones. Lets switch to
SIP phones. That’s a technology we have here and now. I would suggest one
thing. There is a lot in common between India and China and I don’t know
why the countries don’t cooperate in setting some of these standards. The
US standards are very much set in the context of the current environment
where they have a huge legacy to protect here and the people making the
decisions are making legacy decisions. Let me take an example. The whole
trend towards ATM in this country in all of the 90s was driven not
economics, not by the technology, but by what the RBOC needed to do to
preserve their business model. Why India should suffer when they can start
essentially from a clean slate I don’t know. You know we have 2%
penetration of phones. Let that not drive the 98% that we need to put in.
That’s backwards.
So let me suggest that the miracle is right here with 5 billion dollars at
12% tax on the current telecom infrastructure, we have a reasonable hope of
getting there. Entrepreneurial energy is the new oil and we have the human
capital and the energy. We are using our natural resources and that’s
people- we have plenty of those. The internet provides us with worldwide
bazaars and instead of foundation models we have these micro economic
models. Local ROI in every little venture, every individual chooses to do.
I only suggest that people here study the model of the Grameen bank as a
lending institution and say how can we replicate that in the
telecommunications space. A 10x improvement in costs and performance is
easily possible within the next 5 years. I talked about increasing return
economics and I think we can start this “virtuous” cycle. All the
ingredients are there.
Let me stop there and answer questions.
K. Rekhi: Maybe I slept during some part of the talk, but you are not
suggesting a return to planning commissions environment are you?
Khosla: No! We actually already have the notion of blocks. There’s about
5000 blocks in India. I am saying lets stop trying to get
telecommunications to villages and put more where these blocks are.
Rekhi: I heard a lot of “social engineering” in this.
Khosla: No, I am saying lets just put the honey. The social engineering
happens because people behave in their own self interest and market forces
come into play.
Rekhi: I would have personally preferred “No DoT”, “No Planning, none
whatsoever”. Let the entrepreneurs loose. I don’t even believe in having
universal access.
Khosla: There’s 2 models to the world. I started with one which said that
there’s a reality there today. Till you address the needs of 500 million
people on a relatively equal basis the lower bottom half of the scale, you
cant make change happen. You cant avoid the fact that it is a political
democracy in India and not a dictatorship and that politics is an important
driver. I recognize the fact that I can come up with a better model if I
can tell the people exactly what to do.
Rekhi: That’s where I have a problem. Why do we need a model? Let the
market decide.
Khosla: This comes to a very important issue we don’t have enough to
discuss. Let me say that there has been much debate in the US over the need
or no need for an industrial policy. Various administrations go through
deciding whether they have an industrial policy or not. Let me only suggest
that not having a industrial policy is a conscious decision in industrial
policy. Everything you do is a policy decision. You can choose to guide it
or not to guide it. I happen to believe that you guide it towards your
social goals and people vote here like we did and they pick their social
objectives and that then becomes the basis of an industrial policy. If you
are trying to meet the will of the people you have to be guided by those
goals. I can sit in my office and design a set of goals for the country,
that’s pretty easy.
Rekhi: Is this available at the website?
Khosla: I will get it up there Monday or Tuesday.
Jhunjhunwala: Vinod, a number of things that I find very interesting in
your presentation. I disagree with some of you thoughts. Let me start by
saying that first of all 5000 is a good number, we have about 5000 small
towns in the country.
Khosla: Lets not focus on the details and frankly I am not qualified to
pick out whether there are 5000 or 10000 such circles.
Jhunjhunwala: No, No, there are 5000 small towns. The number of class1
cities are only about 300, but if you look at the towns there are about
5000 of them and 6000 railway stations. The fiber as far as I know is going
to about 2000 of these towns and getting it to these 5000 towns is just no
problem. I was disturbed when you said one T1 for a 100,000 people. I was a
little worried about that. I think that was a slip. Fiber is going there
and bandwidth will be available. The point is however that over the last
so many years we have seen about 10-15 cities where a lot of growth is
taking place and maybe another 50-60 where some small growth is happening
and beyond that the small towns have been dying, dying and dying. Therefore
the way I look at it , if we can really strengthen in the country, bring
facilities and education there, yes people can move up and down and if you
get 5000 towns connected then reaching every village is no problem because
wireless will reach every village. I think that is a very good idea. Can we
really put our focus on the 5000 small towns and not on the top 50, which
will grow on their own probably in a bad way today, until we take care of
the 5000 small towns. I totally agree with that. There is no reason why the
whole service industry in India must be in banglaore and Chennai and Bombay
in India and not in these 5000 towns. These could become the design and
educational centers for India. I think India will definitely change if we
pursue this.
Khosla: I only mentioned T1s because one satellite’s IP connections can
serve the whole country. One satellite Ok. If we want to connect 5000
places in say the next 18 months, it can be done for a relatively low cost
Jhunjhunwala: It can be done sooner and cheaper if the DoT were throw its
fiber open to everyone.
S Patil: Lets move on from that.
Jhunjhunwala: I want some real work done on these 5000 centers. If you
starve them by giving only T1s to them you are actually declaring them
Khosla: The point I was trying to make was by deciding the communications
infrastructure, the work centers, the congregation points, we can socially
engineer the whole country and not by dictating anything, but making it
socially attractive for people to move to these locations. The thing that
concerns me the most is it is extremely expensive for people in villages to
relocate to major cities. It is expensive and disruptive. The social costs
are very very high. We want to avoid that and have people stay within their
local neighborhood. You know, 20 kms, 30 kms as I call it, a bicycle
commute, that’s one important design goal. The second is to have enough
centers that you can fund them to be reasonably economic, where you can get
health care and communication, education, access to internet and telephone.
Frankly I don’t think there is any difference between access to internet
and telephones. Looking forward, separating those two modes is silly. I was
saying lets decide what we are trying to engineer for to solve the
fundamental social problem. The issue of trying to catching up with the US
with the PSTN network is a derivative thing and we start looking beyond
that at what is the fundamental thing we are trying to do.
Anne Lee Saxenian: I find this very interesting. The thing I would suggest
is that if you are going to talk about this as social engineering you are
going to have trouble. I wouldn’t use that word.
Khosla: yes I know.
AL Saxenian: We cant plan uses of our capital. If we cant even plan our use
of capital, how can we plan our social world.
Khosla: The term I want to use is lets put the honey pots in the right
AL Saxenian: I think it might be better if we used this model only to plan
for telecommunication or education. There is a bigger message that we can
plan a better social world. That’s a problem and it may be that people
don’t want to move even to these centers.
Khosla: But they don’t have to. They are a bicycle commute away.
Ghosh: You got a response from Kanwal first of doing away with the DoT and
saving the country. I am a little confused that you saying that we have a
system of taxing 10-12% of the revenue of this sector and find some way of
ploughing it back into the sector. Who is going to plough it back? Is it
the players already there? Is it the incumbent with a certain monopoly
directed to do it. This means that there is a planning which will take care
of what you are saying. Secondly when you talk about 5000 circles, it comes
nearly to the 6000 blocks we have in the country. If you are covering a
radius of say even 5-10 kms, you are providing universal access anyway.
Khosla: Yes effectively you reach it. Let me give you a set of statistics
in the US. When we were looking at a cable client being redone. 85% of the
cost of a new cable client in the US is in the last 500 yards. 85% of the
cost!! What I am suggesting is lets not solve the last mile problem. The
cost of getting it to every village is an uneconomical use of our
resources. In that case let bicycle be the communication mechanism for the
last 10-20 kms. If over time if these are the honey pots, the bees will
move to that vicinity and you will get a natural structure that is more
balanced than where we are headed today. As for the issue of monopoly, I
don’t believe in monopolies and there I agree with Kanwal, in that. The
conduit to be built for this should be shared and done cooperatively, and
we have consortiums in US all the time. ATT is putting together a new
consortium to get a new fiber network. Then should the use of the conduit
be limited to one party? In fact you want competition and dynamic change
and that’s where I talked about separating the dynamic pieces from the
static. It is not the government’s job to do this, but its the government’s
job to have an industrial policy that encourages this kind of behavior.
JP Singh: I have a comment and a question. The comment comes from the
social engineering vs laissez faire controversy and I am reminded of the
Punjabi phrase “Paed ginne hain ke aam khaade hain”, which literally
translated means do you want to count to eat the mangoes or do you want to
count the trees and I think I saw you trying to do both here. And somewhere
along the way we have to decide because entrepreneurship is all about this
equilibrium in some ways. My question is this. I was fascinated by your
talk about entrepreneurship and you talked about silicon valley being a
cultural mindset. I didnt know if you could expand on that. But in your
mind what is that cultural mindset, where do we find it in India and what
would be the difficulty of translating it in broad terms and what would it
Khosla: For years, the last 5 to 7 years, people have believed that you
cannot go beyond 200 kms in OC48 optical signaling. The first time somebody
went 2000 kms, within a year there were half a dozen people who claimed
that they could go more than that. Now it is at 40000 km. The very fact
that you know its possible changes your psychology. Entrepreneurship relies
on the fact that once you know its possible, everybody else tries to do it
and in fact if they approach it with the view that its possible they create
all kinds of innovative solutions. We see this repeatedly and that’s where
harnessing the entrepreneurial energy pays off. Then you have competition
and beating each other. 15 years ago you had people claiming that CMOS
would never go beyond 100MHz or that bit densities in magnetic storage
would never cross a threshold. We are now 10 times above those numbers sand
there is no stopping. Once we have crossed the barrier there is no stopping
the entrepreneurs and that’s where the real power of entrepreneurial energy
lies. Enabling this kind of environment is essential. It is seen in the
software business in India. I get emails all the time, somebody saying,
well I didn’t go to college, but I have taught myself Java. Or I did
English at St. Stephens and now I know Java. Why because you are enabling
these people with motivation and drive and role models. I didn’t talk about
role models, but those are extremely important. And people like
Narayanmurthi have broken that mold. They have proved that its possible and
there are 100,000 people trying to copy that. That energetic environment is
the resource you need to transform the country. I don’t know if I have
answered your question. Rafiq, tell me when you want me to stop.
Rafiq: That’s fine. I have question of my own. (laughter). How does one get
the private sector spend the money? Let the govt collect the taxes.
Khosla: You know we need to have nice clear simple policies and
predictability. The private sector then gets the job done. We have seen
this repeatedly. How is the private sector investing in software? Yeah we
can look at all the people losing money in wireless. But what is wrong with
that? They placed uneconomic bets and are paying the price. It doesn’t have
to be a smooth process. We have to acknowledge that entrepreneurship and
experimentation will involve failure. The advantage is that lots of people
are trying and so some experiments will succeed and make 10 times their
money and make up for a lot of those failures at a socio economic level.
It is a much better deployment of resources than having a few carefully
managed projects which is what Kanwal was saying. We just cannot predict
the future.
Rafiq: I would like to support the idea because in the late 80s I was
involved in anevent where a certain migrant community, which was spread
over 900 Gujarati villages, was asked whether they would like to migrate to
what we called 20 centers of opportunity. If they did the only facility
they were given was access to some credit and 6 months of rental support.
What we found was that 70-80% of those asked actually moved, about 20%
moved back. The life of the others was changes forever.
Khosla: This is the whole point. I didn’t plan how much time it would take.
One of the slides talked about “want vs should”. The should method doesn’t
work. “We should have family planning or we should do this.” Want is where
you create the honey pots. I retract the term social engineering. I never
said that. Put the honey pots there and let people do what they would
naturally do. Let them draw the agenda and that’s where entrepreneurial
energy and what Kanwal talks about is the right approach.
Rekhi: Can I ask one more. It is about the notion of honey pots. A lot of
flies will gather near the honey pots. Let the entrepreneur take the risk
and reap the rewards with no preordaining of any rewards. Because Indian
bureaucrats and politicians will jump at the chance of reviving the
planning commission as that’s what they have been doing for 50 years. We
have them on the hook and lets not take them off it right now.(laughter)
Rafiq: We have to make this the last question.
AL Saxenian: It seems to me that we have to face the facts. 35% of all the
venture capital invested in US comes to one small region here. Places all
over the world are trying to imitate this and not succeeding. We have to
think a little bit more about the institutions before all this works. We
cant just say that put a honey pot and VC will come. I think the US govt
spent a lot of money early on in creating infrastructure here. Our
educational system spent a lot of resources training people, part of the
reason we need all the Indians and Chinese is that we don’t train enough
and so need all the H1 visas. I think we have to push the notion of what
makes us Silicon Valley than just saying that it’s a mind set. Providing
capital is the key and you know it very well. VC is distinctive here and in
the rest of the world what is called VC is states providing money or big
old companies providing money with visible assets as collateral. So it is
starting in parts of India, but I am not sure it is enough.
Khosla: Again, I am not pretending that I have solved the countries
problems. I am suggesting a few directions for investigation, more than
saying that these are the right things to do. These arte some principles
and questions we need to look at differently and I am saying that if you
are going to spend a certain amount of dollars and I am sure that we will
spend more than 5 billion dollars in the country in the next 10 years, lets
use a set guiding principles that are different from what we have done. So
we will put it to more economic use. This has nothing to do with venture
AL Saxenian: But entrepreneurship needs microcredit. You mentioned the
Grameen Bank. But small amounts of credit to keep you running is different
from what you normally see in the financial world.
Khosla: I am not trying to replicate the US model and in fact it’s a
mistake to try and replicate this model . The assumptions are different. I
will also disagree with you that the US govt had anything to do with the
Silicon valley phenomenon. They did nothing. In fact they tried to engineer
this in other places. When they had discussions, in the early 90s, I was
part of those discussions – where to put the super computing facilities
AL Saxenian: Maybe in the 50s when you weren’t here.
Khosla: If you look at the history of the valley, the only authentic thing
that I have seen about how and why Silicon valley came to be was Dean
Terman at Stanford Engineering had some ideas, talked about and write some
papers about this industry/university nexus. You can say that it was one
man’s idea, but the US govt and the county had nothing to do with this
HH: What about the defense contracts.
Khosla: Well I am sorry, I should stop here. I am happy to take these
discussions offline or on e-mail. Like I said in the beginning I am sure
there are lots of holes. Thank you everybody.

Written by Bhushan Kulkarni

December 28, 2006 at 11:00 am

Posted in Management

Tagged with ,

Leave a Reply

Please log in using one of these methods to post your comment: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: