Abstract: We have had the great fortune of working in a profoundly and broadly impactful discipline during its formative years, learning our craft from some of its giants. While I am far less adapt at seeing and shaping the future than most of you, I can bear witness to what it takes for a technology to successfully impact business. In this talk, I will share insights from 25 years of entrepreneurship.
CMU CSD. At the risk of opening with one of those “I used to have to walk through a mile of snow”, when I arrived at CMU in the Fall of 1980, I’d barely done any programming at all. My high school had one IBM computer and you programmed it with punch cards! Even at CMU there were still terminals with rolls of paper called DECwriters rather than screens—imagine trying to program by typing in the line number of the code you wanted to change, or worse shuffling through a deck of cards?
In those days, we programmed in the middle of the night, because that was the only way to get sufficient cycles that a build wouldn’t bring a shared machine to its knees. Most of us would sleep through class (or in class) in order to get the vastly higher productivity we could get by coding in the middle of the night. I remember it was novel not to know quite how many VAX or PDP-11 computers CMU CS had at a given point in time. Groking Moore’s Law never prepared me for the computing horsepower we have at our disposal today.
My CMU acceptance said school of engineering, but all my energy went into Computer Science. My first three courses were Bill Scherlis’ 211, where I was introduced to the science of computing, to my first real programming in Alfred Spectors’ 212 to Comparative Programming Languages with Guy Steele, where we coded up solutions to the same problem in languages like SNOBOL and APL. When I was first learning about AI in one of Geoff Hinton’s first classes, I remember a talk from Herb Simon that it was going to be easier to teach a computer to be a professor than to drive a jitney, because evolution had been working on hand/eye coordination much longer than it had on abstract thought. Guessing that bet goes the other way, but hard for me to imagine a better on ramp. I was hooked, and was lucky enough to have gotten accepted to stay on for my Ph.D. So I turned down my job offer in ’84 from some obscure micro-computer upstart named Microsoft that I never thought would amount to much to remain in CSD until I ran out of degrees.
And for a small town kid, Pittsburgh was actually a pretty awesome place to come. You could still smell the steel mills in the early morning back then, but there was a bar called the Decade just down in Oakland that used to get some crazy good acts, including the Pretenders, U2, and the Red Hot Chili Peppers. I was underage and missed those, but do you know the Clash played on CMU’s campus? And I got to drink scotch with Kurt Vonnegut and one of my coconspirators in music Josh Bloch got to drink Wild Turkey with Hunter S. Thompson. I have many fond memories of those times.
When I got my bachelor’s degree, I was a bit cocky. I considered myself to be in the top handful of undergrads coming out of CMU, but was profoundly humbled by the caliber of my fellow grad students and faculty. That was when reality started to dawn on me that life as the faculty member I aspired to be was going to be a major leap. I’d enjoyed the success I did because I was working with hugely talented people showing me the way. Without Frank Pfenning, I wouldn’t have my Ph.D. Thanks Frank. Frank would intuitively know when I was stuck and gently point me in a more promising direction.
It was weighing life post Ph.D. that two other CMUers came to my rescue. First there was a talk by Bruce Nelson on What They Don’t Teach You in Grad School, all about becoming a successful entrepreneur. That talk inspired me. My dad was an entrepreneur, and I figured doing so in tech would let me straddle software and business. In addition to Frank, my other deepest debt is to Alfred Spector, who invited me to Transarc and served as a mentor, seeing in this failed researcher the makings of an entrepreneur. Without that bet, I’m quite sure I would not have enjoyed the great career luck I have had.
Entrepreneurship. Over the last twenty five years, I have had the great fortune of working at four startups—Transarc, WebLogic, Zimbra and Pure Storage. (I never liked the term serial entrepreneur, as it reminds me of the other serial—killer.) I am drawn to startups because I love the opportunity to help create something new out of whole cloth, to help craft the product and business model to fix something broken, to unlock a market and to build something great. Tech entrepreneurship is an ideal place to be if you want to have one foot in technology and another in business.
Top talent also gravitates to startups because we want to shed legacy constraints. (One of my standard recruiting lines: “Do you really want to squander your talent turning 15 year-old software into 20 year old software?”) The other reason I keep ending up at start-ups is I hate internal meetings. We have a policy at Pure that any employee trapped in a soul-sucking status meeting has CEO prerogative just to walk out. (That really happens by the way—I’ve had managers complain that I told employees to bail from meetings. Fix your meetings was my response.) While no doubt start-ups can be lucrative, in my experience that is not the right motivation: Be a missionary seeking to build something great rather than a mercenary.
The common thread between all of four start-ups I’ve been part of is now my recipe for entrepreneurial success:
- Choose a big market, because you want to have a material impact
- That is ill served by current vendors and solutions
- And better yet, is experiencing a technology sea change (such as from mechanical disk to flash memory) that further disadvantages the incumbents
- Then define a disruptive recipe with sustainable advantage, ideally both technology and business model (it can be very hard for incumbents to respond to new business models)
- And of course build a hugely gifted team
It is often said that start-ups inevitably tack to some new market opportunity. That hasn’t been true in my experience—we stuck to our original vision.
Edison said genius is one percent inspiration, ninety-nine percent perspiration. No doubt research requires a lot of sweat, but I think the ratio may be worse for entrepreneurs. Lesson #1 is that to be commercial viable, technology has to be extremely easy to use and highly reliable. Complicated or fragile products will never get mass adoption, but making complicated things simple is really hard. There is still a lot of mediocre software in the world. We can do better.
Lesson #2 is it is also really hard to get the market’s attention. Building a better mousetrap is necessary but not sufficient, because enduser time and money are scarce! Think about how grateful you are for each new email solicitation. You need to deeply understand enduser pain and aspirations to break through. The most typical downfall for an entrepreneur is cool tech, but insufficient value to command enduser attention.
Lesson #3 is once you open the door, then you need to get to the predictable and repeatable sell. That is, you need to be able to easily identify a target customer, the likelihood they will buy, and be able to prove out their path to value. Start ups often flail and exhaust their cash because they start growing sales and marketing before they hone this rinse and repeat.
My favorite quote on the essence of entrepreneurship is stolen from the military. Admiral Jim Stockdale once said
You must never confuse faith that you will prevail in the end—which you can never afford to lose—with the discipline to confront the most brutal facts of your current reality, whatever they may be.
Like great research, successful entrepreneurship stretches you. You have to stay focused on the long-term vision—to see the path through to success. But you also have to stay grounded. Right ratio is that you spend 4 days a week thinking your going to hit it out of the park, and 1 day a week thinking your crashing and burning. Unfortunately, you will likely go on for years in that state before you know for sure, but if the crash and burn days become the majority, you should think hard about what to change. The hardest part of entrepreneurship for me is assuring the successful outcome for the team and for investors—you think hard about that before taking on another start-up.
Looking forward. For computer science, I think these are the good old days. I didn’t have my recursion epiphany until I was a sophomore. My son, a budding Scratch programmer at 8, may get there in elementary school. Technology is so much more accessible today than it was back then. In retrospect, we take them so much for granted that it is easy to miss the simple elegance and dramatic impact of the spreadsheet, or the hyperlink, or the touch-screen. Again I am still struck by the image of my son, then three, poking the TV screen expecting a response, and expressing frustration that it must be broken.
It is fun to also contemplate that our discipline has become far more popular than it used to be. Being a geek used to carry a stigma in part because computers were so inaccessible back in the day. While in Pittsburgh on the rare occasions I had the opportunity to meet a girl at a party (women in our discipline is another thing that’s gotten dramatically better since the old days), I would say I was an underwater welder rather than a programmer. I don’t recall it ever actually working, but the point is today we have HBO shows and movies aggrandizing tech. Who would have thought being a computer nerd would ever be cool?
When I last returned to CMU for a celebration of 50 years of innovation in computing, I was startled to hear voiced the worry that our discipline was getting boring—that perhaps all of the revolutionary work had been done. Hopefully recent events have helped to dispel this silliness. Our industry is awash in opportunity. Mobility, cloud, open source, and so on have dramatically lowered the hurdles to innovation and are causing a holistic rethink of the tech industry:
- The business world is poised to repeat their version of the consumer-tech revolution that yielded Google, Apple and Facebook. I think this is a great time to launch new programming languages, new database designs, and new application platforms.
- As Marc Andreessen eloquently put it, software is eating the world. I used to lament the fact that the cell phone industry was driven outside the USA and that they were lousy Internet devices. Thanks Apple and Google for the smart phone. Tech is disrupting industries I would have never expected, like the smart car.
- Many of the long-term aspirations of AI and machine learning are getting knocked off: language translation, expert systems for investment management to medical diagnosis. I harkened back to Herb Simon’s talk when my Tesla did the driving to work this week. With fearless 8 and 4 year olds at home, I’m particularly looking forward to self-driving cars.
- And computation continues to revolutionize other disciplines – Astro- and particle physics and genetics are all big data problems. At his lab at UCSF, Joe Derisi is doing full spectrum DNA sequencing of spinal fluid, blood and lung tissue, spotting esoteric diseases and autoimmune disorders missed by symptomatic diagnoses.
Let me close with a story: Back in the early days of Java, I had the opportunity to represent WebLogic in a back-room debate among the Java powers that be. I made what I thought was a well-reasoned argument. My case was summarily dismissed by one of the Sun representatives, who closed his response with “And Scott, just why in the hell would you think that would ever work?”. I paused and responded with a mantra prescribed by Rich Sanzi, a fellow alum, “I guess I’m just a blind, stupid optimist.” At that point, one of my allies in the debate said laughing, “Geez Scott, I sure hope you don’t lose your optimism, because then you’re just going to be blind and stupid!”
There’s an underlying truth in this story: If you aspire to lead people or an industry, undying contagious optimism is an essential asset. Cultivate it. Just try not to be blind and stupid.
Thanks to your efforts and those of our industry, the barriers to innovation in Computer Science continue to drop. You just need the spark.
Thank you this opportunity, both tonight and the one afforded me 30 years ago. Then and now I am sure my arrival lowered the average IQ of the place. I am profoundly grateful for all that the CMU Computer Science Department has done for me. My career successes grew out of the seeds planted here. I can’t wait to see what the next fifty years will bring. Thank you.
