Vinod Khosla, leading entrepreneur, venture capitalist and billionaire on reinventing societal infrastructure, nurturing the culture of innovation in large companies and what it takes to be a transformational CIO.
WOOL: Vinod, thanks for being with us and sharing your ideas and insights. Tell us your ideas and passion about solving the big societal problems, about creating the infrastructure we need, and about bridging the gap between those who have a resource rich lifestyle and those who don't.
VINOD KHOSLA: The lifestyle of the top 10% of the population is rich in nutrition, housing, transportation, healthcare and any other dimension - what you call, resource rich. But if the remaining 90% aspire for the same kind of lifestyle, there just isn't enough of land, water, energy, medical care and other resources to go around. The world just cannot bridge the gap between what we have and what we want, without straining the environment.
The only leverage we have is technology, which can multiply resources 2x to 10x, without straining the ecosystem. Four years back, when I turned 60, I wanted to focus the next 20 years of my life on solving big problems. I analysed almost every sector of the non-governmental GDP and tried to figure out how to apply technology to multiply resources. And I shocked myself by coming up with 2-3 different technologies in every sector that could achieve a non-linear use of resources. Now, if I could do that as an armchair exercise over my August break, I am very optimistic that in the next 30-40 years there will be visionary entrepreneurs who will come up with actual solutions.
WOOL: Which are the areas most ripe for disruption, in your opinion, and why aren't more people taking up these challenges?
VINOD KHOSLA: Almost every area offers the potential for massive disruption that is impossible to predict until someone actually takes it up. Look at examples in retailing - the disruption did not come from traditional retailers but irrational outsiders like Bezos. Similarly, large automotive companies thought that electric cars were years away until an "unreasonable" person like Musk changed those assumptions, and now everyone is following him! Even the changes in the space industry like reusable rocket launch systems, are coming out of SpaceX, not established players. It was hard to predict a few years back that we could launch rockets for 20-30 million USD or 3-D print entire motors; that premium beef burgers could be made from plant protein, using 95% less land and 85% less water! The nature and timing of disruption seems to depend a lot more on irrational pursuits of people who decide to solve a problem, rather than the nature of the problem itself. As to why there aren't more people taking up challenges - I think it's because most people think of themselves as rational human beings. They expect the world to behave in a certain way and don't take up the irrational challenges! The scope of what can be achieved isn't limited by actual potential but by imagination. Take public infrastructure, for instance. Why are we limited by the assumption that we must have a fixed route? A city like San Francisco can reimagine public infrastructure as a point-to-point service and in the process, it can reduce both congestion and crosstown travel time by 30 minutes! In 20 years, we could leverage autonomous vehicles, reserve special lanes for public transport, or do away with private vehicles altogether! My math says it can be done for a dollar a ride!
WOOL:But the venture capital that nourishes these ideas typically doesn't have the patience that these large problems need for finding solutions, so how does that contradiction get resolved?
VINOD KHOSLA: Actually there isn't a contradiction at all. Neither the VCs nor the entrepreneurs have a long term, detailed roadmap of the future, even the wildly successful ones. Take Uber's valuations as a guide for accuracy of estimates. 10 years back one might have imagined it would be half a billion dollars; 7 years back 1.5 to 2 billion and even 3 years back, 10 billion dollars would have been a good estimate! All of these proved to be way off the mark. Uber has evolved through 16 steps and it's likely there are more ahead. I talked about reimagining public transport as a point-to-point service, which could potentially eliminate the need for an Uber-like service altogether. But it's an evolutionary process that, like nature, produces the most beautiful designs to solve problems, which top-down long-range planning cannot do. In large companies, the long-range planning process is almost a sacred ritual but it hasn't helped them reinvent the future at any time. In every market in the last 30 years, not one breakthrough innovation has come from long-term business plans. Media innovations came from Twitter, Facebook and YouTube, electric vehicles were the brainchild of Tesla and driverless cars came from Google and others.
Innovation is and has always been a messy, wild, zig-zag journey of unreasonable entrepreneurs. In 2000, most predictions and planning about telephone lines in India were about landlines, with no attention given to mobile phone technologies, which evolved out of the efforts of a bunch of people, trying out different things. In 1996, most large telecommunication companies in the US and network router companies had no plans to build the public infrastructure based on TCPIP protocol, they were reliant upon the more established ATM protocol. That created an opportunity for Juniper Networks. Even 15 years ago, having had initial success in e-commerce, who would have predicted that Amazon would dominate the corporate computing infrastructure market? Innovation is not based on linear predictions made well in advance. That is an illusion, generated by hindsight.
WOOL:Why do you think large companies or institutions are not successful at breakthrough changes?
VINOD KHOSLA: Entrepreneurs have a vision and they make that vision happen partly because they are tone deaf to cynicism or ridicule. You see, entrepreneurs and institutions live in different worlds, in different paradigms. They live on different parts of the risk-reward curve. Large companies dial down the risk to improve the likelihood of success. But in the process, the consequences of their success become inconsequential! The risk is ironed out of the system, but so are large expected payoffs.
For entrepreneurs, nothing matters like success, so they actually increase the probability of failure, which is risky but chances are that they will end up with significant payoffs when they succeed. And that's why so many start-ups fail, it's the way it's supposed to work. We are tricked into believing that improbable ideas are unimportant. The truth is they are the only ones that matter. Progress is a series of improbable changes - not extrapolations of the past by experts. Experts are right about the future when nothing changes, so their predictions have little value. We get tricked because we can't pick today, which improbable ideas will shape the future but that doesn't mean everything improbable is non-essential.
WOOL: So how can large organizations learn to take risk and build a start-up culture inside? How does one get that DNA inside an institution?
VINOD KHOSLA: Human beings respond to stimulus, and most large companies reward only success. If failure is what feeds future success, large companies must start rewarding intelligent failure, as much as they reward extrapolated success. At the same time, failing ideas and projects must be shut down quickly, which doesn't happen fast enough or often enough in large companies without penalizing the people responsible for trying. This approach to risk and failure needs to change and some companies actually celebrate the "best failure", to take the taboo and stigma out of intelligent failure and risk taking. Take Google, by far its most important product is the search engine. But two years back, it completely rebuilt the search using AI without the end users realizing it. It was a huge risk and by traditional thinking, it wasn't a compulsion to undertake it.
WOOL: But can large organizations be fast followers instead, by adopting winning ideas early and scaling rapidly, using distribution, capital, and other assets?
VINOD KHOSLA: It is not impossible but the evidence is against it. And there are two good reasons. One, success depends on talent, and large organizations do not have access to the kind of talent that's required. A 12-person start-up that depends on brilliant AI engineers, could end up paying a couple of them seven-figure salaries, more than even the CEO, but large organizations can't upset their apple-carts that easily. And the second reason is that as the original idea evolves, it meets with hundreds of failures; each becomes a part of institutional learning as wrinkles are ironed out. An outsider can use experts to understand and emulate two big things about a successful idea, but the internal knowledge born of hundreds of failures is something that cannot be replicated.
Large, resourceful auto companies have tried to create something like Uber's ride sharing service but can't replicate its success because they haven't been exposed to the thousands of mistakes Uber went through. And that's also true for Amazon in ecommerce and cloud computing, and Google in search.
WOOL:Are there inherent limitations to institutions since several large companies are publicly listed and have short term commitments to the street?
VINOD KHOSLA: Wall Street goes by expectations so if you set expectations of increasing earnings, you will be on that treadmill. Bezos has avoided that by explicitly aiming for growth and the risks that come with it. Elon Musk has done the same, even though everyone might not be comfortable with the financial risk Tesla has taken. But they are true to their belief that risk is what they need to succeed, and that's exactly what they have communicated to the street.
WOOL: What role do CIOs have to play in large organizations? How does one take on the role of a Transformational CIO as opposed to a Run CIO?
VINOD KHOSLA: 20 years back, the tech sector was largely about enterprise software, with a little bit of consumer and mobile and commerce, but it wasn't about most of global GDP. Today tech is about every part of every business, so how technology is managed is the biggest leverage in every company, and Boards are keenly watching that too.
External consultants cannot provide technology strategy because success depends on accumulation of knowledge. The CIOs are the most technologically proficient and best positioned to guide their companies on tech strategies. Take the example of one of our companies - it is large, at USD one billion of revenues, and has 2,000 employees. It can easily grow at 40% per annum for the next four years and can achieve 4X in revenues and employees. But can it get there with half as many employees as the linear model suggests? It needs to look at every area where it can drive dramatic productivity, not by 2% or 5% but 2 to 5 times! The only way to do that is to use AI, which can resolve 80% of support tickets or automate 80% of back-office processing or provide automated suggestions to the sales team for conversations. Who else but the CIO is best placed to drive these changes in the company and think about technology strategically?
WOOL: Large organizations continue to invest and acquire innovative technology companies. How can they get better at integrating and deriving the value they pay for?
VINOD KHOSLA: Culture matters most when it comes to innovation. While Silicon Valley continues to face criticism for its move-fast-and-break-things way of functioning, it will be even more important in the future to have a culture that encourages experiments and failures because the rate of change is escalating rapidly, and the environment is becoming more dynamic. CEOs of many large companies just don't seem to feel comfortable in the zone that produces less predictable success but with much higher payoffs.
All the banks put together have failed to produce the most important changes in small business payments - like Stripe or Square. Despite a lot of publicity about the big banks investing in AI, Analytics and Blockchain, I am sceptical that the next big innovations in banking will come from the established firms.
WOOL: As an investor, can one get better at predicting the disruption in enabling infrastructure on which the next wave of innovative businesses will be built?
VINOD KHOSLA: While directional predictions are possible, the specifics are hard to predict. I describe myself more as someone who speculates about rather than predicts technological change. It's easy to come up with a range of possibilities and be embarrassingly wrong about specifics. In 2004, three years before the arrival of the iPhone, I speculated that the phone would be a computer used for everything except making a phone call. I wrote a piece called, "The device that used to be the phone" but the specifics were embarrassingly wrong. Like many others, I found it hard to imagine a phone without a keyboard. The future may not be predictable, but it is definitely discoverable, and therein lies the ground for optimism. My entrepreneurial career started in 1980 and every decade has seen the landscape evolve dramatically. The axes of innovation are new technologies, which are evolving continuously, and combining in exciting ways. Take CRISPR, AI, 3D printing, Material Science, Robotics and newer ones emerging every year - the large number of axes produce exponentially more possibilities for solving large problems, so even large organizations must keep feelers out at the edge and be ready to respond.
Some changes are around the corner - in another three years, Robotics and 3D printing can upend the economics of Chinese labor for the most basic of tasks and reshape the supply chain for a lot of manufacturing industries. A little farther down the road, a combination of 3D printing, AI and Material Science will reduce the materials in construction to one tenth of traditional requirements and change the construction business.
WOOL: Finally, let's go back to one of the key pillars of societal infrastructure - healthcare. What do you think we can do to accelerate the adoption of breakthrough technologies like AI in that space?
VINOD KHOSLA: I think we need to challenge assumptions about the future of healthcare delivery and we are doing that in one of our start-ups. Instead of focusing on getting the primary care physician to use AI, we are building an AI based primary care physician, and planning to build AI cardiologists, AI psychologists, AI oncologists or even a virtual tumour board. I predict that in 20 years we will have better healthcare available in remote villages in India than we have in Stanford because instead of depending on a highly trained expert, we can deliver sophisticated AI based healthcare through a high school graduate who has been trained on using the platform for a couple of years. And all this for the cost of a Google search!