The Aha! Moments That Made Paul Graham's Y Combinator Possible

When it comes to funding startups, Y Combinator is an incubation machine. And behind it all are founder Paul Graham's unique insights about what constitutes true innovation.

In the summer of 1995, Paul Graham heard a story on the radio promoting the endless possibilities of online commerce, which at the time hardly existed. The promotion came from Netscape, which was trying to drum up interest in its business on the eve of its IPO. The story sounded so promising, yet so vague. At the time, Graham was at a bit of a crossroads. After graduating from Harvard with a PhD in computer science, he had fallen into a pattern: he would find some part-time consulting job in the software business; then, with enough money saved, he would quit the job and devote his time to his real love—art and painting—until the money ran out, and then he would scramble for another job. Now 30 years old, he was getting tired of the pattern, and he hated consulting. The prospect of making a lot of money quickly by developing something for the Internet suddenly seemed very appealing.

He called up his old programming partner from Harvard, Robert Morris, and interested him in the idea of collaborating on their own startup, even though Graham had no clue where they would start or what they would develop; eventually, they decided they would try to write software that would enable a business to generate an online store. Once they were clear about the concept, they had to confront a very large obstacle in their way. In those days, for a program to be popular enough it would have to be written for Windows. As consummate hackers, they loathed everything about Windows and had never bothered to learn how to develop applications for it. They preferred to write in Lisp and have the program run on Unix, the open-source operating system.

They decided to postpone the inevitable and wrote the program for Unix anyway. To translate this later into Windows would be easy, but as they contemplated doing this, they realized the terrible consequences it would lead to—once the program was launched in Windows, they would have to deal with users and perfect the program based on their feedback. This would mean they would be forced to think and program in Windows for months, perhaps years. This was too awful a prospect, and they seriously considered giving up.

Paul Graham

One morning Graham woke up with the idea that they might be able to control the software on the server by clicking on links. He suddenly sat bolt upright, as he realized what these words could mean—the possibility of creating a program to set up an online store that would run on the web server itself. People would download and use it through Netscape, clicking various links on the web page to set it up. This would mean he and Morris would bypass the usual route of writing a program that users would download to their desktop. It would cut out the need ever to have to dabble in Windows. There was nothing out there like this, and yet it seemed like such an obvious solution. In a state of excitement he explained his epiphany to Morris, and they agreed to give it a try. Within a few days they finished the first version, and it functioned beautifully. Clearly, the concept of a web application would work.

Over the next few weeks they refined the software, and found their own angel investor who put up an initial $10,000 for a 10 percent share in the business. In the beginning, it was quite hard to interest merchants in the concept. Their application server provider was the very first Internet-run program for starting a business, at the very frontier of online commerce. Slowly, however, it began to take off.

As it panned out, the novelty of their idea, which Graham and Morris had come upon largely because of their distaste for Windows, proved to have all kinds of unforeseen advantages. Working directly on the Internet, they could generate a continuous stream of new releases of the software and test them right away. They could interact directly with consumers, getting instant feedback on their program and improving it in days rather than the months it could take with desktop software. With no experience running a business, they did not think to hire salespeople to do the pitching; instead, they made the phone calls to potential clients themselves. But as they were the de facto salespeople, they were also the first to hear complaints or suggestions from consumers, and this gave them a real feel for the program’s weaknesses and how to improve it. Because it was so unique and came out of left field, they had no competitors to worry about; nobody could steal the idea because they were the only ones who were insane enough to attempt it.

Naturally, they made several mistakes along the way, but the idea was too strong to fail (though Graham notes the company came close several times anyway); and in 1998 they sold their company, named Viaweb, to Yahoo! for $50 million.

As the years went by and Graham looked back at the experience, he was struck by the process he and Morris had gone through. It reminded him of so many other inventions in history, such as microcomputers. The microprocessors that made the microcomputer possible had originally been developed to run traffic lights and vending machines. They had never been intended to power computers. The first entrepreneurs to attempt this were laughed at; the computers they had created looked hardly worthy of the name—they were so small and could do so little. But they caught on with just enough people for whom they saved time, and slowly, the idea took off. The same story had occurred with transistors, which in the 1930s and ’40s were developed and used in electronics for the military. It was not until the early 1950s that several individuals had the idea of applying this technology to transistor radios for the public, soon hitting upon what would become the most popular electronic device in history.

What was interesting in all of these cases was the peculiar process that led to these inventions: generally, the inventors had a chance encounter with the available technology; then the idea would come to them that this technology could be used for other purposes; and finally they would try out different prototypes until the right one fell into place. What allows for this process is the willingness of the inventor to look at everyday things in a different light and to imagine new uses for them. For people who are stuck in rigid ways of seeing, the familiarity of an old application hypnotizes them into not seeing its other possibilities. What it all really comes down to is the possession of a flexible, adaptable mind—something that is often enough to separate a successful inventor or entrepreneur from the rest of the crowd.

After cashing in on Viaweb, Graham hit upon the idea of writing essays for the Internet—his rather peculiar form of blogging. These essays made him a celebrity among young hackers and programmers everywhere. In 2005 he was invited by undergraduates in the computer science department at Harvard to give a talk. Instead of boring them and himself by analyzing various programming languages, he decided to discuss the idea of technology startups themselves—why some work, why some fail. The talk was so successful, and Graham’s ideas so illuminating, that the students began to besiege him with questions about their own startup ideas. As he listened, he could sense that some of their concepts were not far off the mark, but that they badly needed shaping and guidance.

Graham had always intended to try his hand at investing in other people’s ideas. He had been the beneficiary of an angel investor in his project, and it was only right to return the favor by helping others. The problem was where to begin. Most angel investors had some related experience before they began investing, and they tended to start out on a small scale to get their feet wet. Graham had no such business experience. Based on this weakness, he hit upon an idea that at first glance seemed ridiculous—he would synchronously invest $15,000 in ten startups all at once. He would find these ten prospects by advertising his offer and choosing the best among the applicants. Over the course of a few months he would shepherd these novices and help guide them to the point of launching their idea. For this he would take 10 percent from any successful startup. It would be like an apprenticeship system for tech founders, but it really had another purpose—it would serve as a crash course for him in the investing business. He would be a lousy first investor and his pupils would be lousy entrepreneurs, making them a perfect match.

Yet again he recruited Robert Morris to join him in the business. A couple of weeks into the training, however, he and Morris realized that they were actually on to something powerful. Because of their experience with Viaweb they were able to give clear and effective advice. The startup ideas they were shepherding looked quite promising. Perhaps this system they had adopted as a way to learn quickly was an interesting model in itself. Most investors only handle a few startups a year; they are too overwhelmed with their own businesses to handle much more. But what if Graham and founding partner Jessica Livingston were to devote their time exclusively to this apprenticeship system? They could mass-produce the service. They could fund hundreds instead of dozens of such startups. In the process they would learn in leaps and bounds, and this exponentially increasing knowledge would lead to increasing numbers of successful startups.

If it really took off, not only would they make a fortune, but they would also have a decided impact on the economy, unleashing into the system thousands of savvy entrepreneurs. They called their new company Y Combinator and considered it their ultimate hack to change the shape of the world’s economy.

They coached their apprentices in all of the principles they had learned along the way—the benefit in looking for new applications of existing technology and needs that are not being met; the importance of maintaining the closest possible relationship with customers; the need to keep ideas as simple and realistic as possible; the value of creating a superior product and of winning through craftsmanship, as opposed to fixating on making money.

As their apprentices learned, they learned as well. Oddly enough, they discovered that what really makes successful entrepreneurs is not the nature of their idea, or the university they went to, but their actual character—their willingness to adapt their idea and take advantage of possibilities they had not first imagined. This is precisely the trait—fluidity of mind—that Graham had identified in himself and in other inventors. The other essential character trait was supreme tenacity.

Over the years, evolving in its own way, Y Combinator has continued to grow at an astounding rate. It is valued now at $500 million [editor's note: Graham puts Y Combinator's current assets at less than $500 million], with the clear potential for further growth.

The lesson is simple—what constitutes true creativity is the openness and adaptability of our spirit. When we see or experience something we must be able to look at it from several angles, to see other possibilities beyond the obvious ones. We imagine that the objects around us can be used and co-opted for different purposes. We do not hold on to our original idea out of sheer stubbornness, or because our ego is tied up with its rightness. Instead, we move with what presents itself to us in the moment, exploring and exploiting different branches and contingencies. We thus manage to turn feathers into flying material. The difference then is not in some initial creative power of the brain, but in how we look at the world and the fluidity with which we can reframe what we see. Creativity and adaptability are inseparable.

This excerpt is from Mastery by Robert Greene, reprinted with permission from Viking Press. The web version of this article has been updated to reflect corrections and clarifications noted by Paul Graham on Y Combinator.

Robert Greene is the author of the international bestsellers The 48 Laws of Power, The Art of Seduction, The 33 Strategies of War, and The 50th Law. His highly anticipated fifth book, Mastery, examines the lives of great historical figures such as Charles Darwin, Mozart, Paul Graham, and Henry Ford and distills the traits and universal ingredients that made them masters. He has a strong following within the business world and a deep following in Washington, DC.

[Image: Flickr user Kim Love]

Add New Comment

6 Comments

  • Mike

    While Y Combinator is a great deal better than going through the vile VC system, it is still based on the same flaw as VC's.

    It is also why no Y Combinator or VC funded technical business has gone very big.

    Whether the you go though YC or VC, the goal is the same: to get bought out.

    There is no goal to produce a fantastic product that pushes the bounderies. No goal to build up a stable, long lasting business. The goal is to sell out as quick as possible, which does not lend itself to building a stable company that creates solid, innovative software.

    The current system exists soley for making money which is why tech companies that go the YC or VC route go nowhere.

    Of course, there are more benefits to YC, because you get access to a lot of experienced people, so maybe creating a throw away company just to get all that training is worthwhile. Then, the person can create something of substance with the money they made selling out.

    Viaweb went nowhere under the unskilled direction on Yahoo!, in fact it was made worse by switching to C++ instead of keeping the language it was originally written in: Lisp. Lisp is more expressive and far less verbose than C++. Many things that are possible in Lisp are either impossible or much more difficult. This led to ViaWeb losing features and reliability. C++ programmers are cheaper and more abundant than Lips Programmers, however the money wasted in C++ programmers making Viaweb less functional could have easily hired 2 Lisp programmers that could have managed the entire system.

  • Wrenj

    If only Paul and friends would take on other "verticals". It would be like Steve Jobs taking Apple's legendary skills into other markets.
     
    Who else is better equipped to disrupt the startup "market" for things other than software-based businesses?

  • Allan Hytowitz

    Fascinating article.  

    As the inventor and developer of the patented Dyop™ test for acuity (see www.dyop.org), it is reassuring that I am on the right track towards commercializing my disruptive technology as to better quantifying how we really see.