Stephen Wolfram doesn’t blog much, but when he does he makes it count. On November 13, 2013, Wolfram sat down at his Mac and promised that the new computer language he created would be his “most important technology project yet.”
All bloggers are in the business of self-promotion. But those were particularly strong words coming from a man who has already invented many of the underpinnings of today’s revolution in technical computing. Indeed, in his lengthy blog post announcing the “Wolfram Language,” Wolfram produced what Paul Kedrosky of the Kauffman Foundation called “one of the most entertainingly hubristic product semi-announcements ever.” But Wolfram doesn’t do small. As an inventor and businessman, he has a history of making big bets—and covering them. His announcement may have been a boast, but it was not an empty one.
The 55-year-old Wolfram is best known for his technical achievements. He’s the founder, president, and CEO of the privately held Wolfram Research. The company publishes his computational software program Mathematica, which has become the standard for technical computing and is broadly deployed across scientific, engineering, mathematical, and computing fields. Mathematica, the program that made Stephen Wolfram’s fortune, is the VisiCalc of math programs. It changed the way scientists and mathematicians use math on their computers in many of the same ways that spreadsheets helped expand how people could manage businesses—making mathematical work on PCs both more powerful and easier. But, significantly, unlike VisiCalc, Mathematica has not been rendered redundant by its descendants. More than 25 years after its introduction, it continues to be a standard, particularly in academia. It remains a rich, open-ended environment in which a vibrant ecosystem of add-ons and consultants flourishes.
Wolfram Research also publishes Wolfram|Alpha, a computational knowledge engine that directly addresses factual inquiries by computing the answer from a range of external resources, rather than providing a list of links that might contain the answer, as you would expect from a standard search engine. Wolfram|Alpha serves as the “factual” back end to Apple’s Siri (as well as part of Microsoft’s Bing and the independent search engine DuckDuckGo). It powers a personal analytics engine for Facebook and is embedded in The Elements, one of the first apps that showed what an iPad could do. It includes manifold sources, such as the CIA’s World Factbook and the CrunchBase database of technology companies, and is so large that it runs across more than 10,000 CPUs. Some of the data sets are automatically generated, while others are curated and corrected by humans. The end result, written in more than 15 million lines of Mathematica code, helps people find unexpected, and sometimes powerful, connections across multiple databases.
Wolfram Research, which today has about 400 employees, has thrived for 26 years (Mathematica and its extensions remain the core business of the company), even when Wolfram himself went on a nearly decade-long semi-sabbatical to research and write a 1,192-page book not-so-humbly titled A New Kind of Science (Wolfram Media, 2002). (He characterized himself as an engaged but emphatically part-time CEO during that time.) The book, which has spawned many high-profile and energetic advocates—and no small number of critics—argues that the simplest of programs underlie our most complex universes. When Wolfram was looking for luminaries to write blurbs for A New Kind of Science, his longtime friend Steve Jobs advised him against that approach, in typical Jobsian terms. “Isaac Newton didn’t have back-cover quotes,” Jobs told Wolfram. “Why do you want them?”
The arguments in A New Kind of Science are profoundly iconoclastic and aim to change the world at an Isaac Newton level. The basic premise is that simple rules that work like basic computer programs can lead very quickly to surprising complexity. If our traditional understanding of science comes primarily from engineering and mathematics, Wolfram’s argument revolves around the idea that computing can explain more about the complexity of our world than those two disciplines alone. Indeed, he seems to believe that thinking of everything as a computer program is the path to understanding. He questions some basics of modern science, arguing, for example, that natural selection is not the primary cause of complexity in biology, and that the second law of thermodynamics might have an exception. From Wolfram’s vantage point, computing isn’t just a way to learn more about engineering and mathematics; it’s a fundamental and new way of looking at how science works—and how our world works. The initial work Wolfram did on the topic has certainly entered into the scientific mainstream—his research in the 1980s in cellular automata, a once-obscure branch of physics, has been cited in more than 10,000 academic papers—but it’s as yet unclear whether those mostly accepted notions can be extrapolated to explain what Douglas Adams once called “life, the universe, and everything.” Everything, A New Kind of Science argues, from the cellular level to the whole universe, runs like a computer program, deriving increasing complexity from the simplest of rules. By looking at the universe that way, he says, we can comprehend it as it really is.
Regardless of how Wolfram’s arguments are or aren’t integrated into the scientific canon over the decades to come, what many outside the academy find most impressive about A New Kind of Science is its audacity. Wolfram performed his work more or less independently, outside the mainstream of scientific research; he published the volume himself; and his work has been covered by the general press with an intensity that many serious scientists might have trouble imagining. Yet the tone of the book can be off-putting: Chris Lavers, in the Guardian, characterized A New Kind of Science as “the most arrogant piece of science writing I have ever read.” Indeed, Wolfram’s book starts off at full speed, aiming to crash through any accusations of small thinking: “Three centuries ago science was transformed by the dramatic new idea that rules based on mathematical equations could be used to describe the natural world. My purpose in this book is to initiate another such transformation, and to introduce a new kind of science that is based on the much more general types of rules that can be embodied in simple computer programs.”
Many critiques of A New Kind of Science have focused on Wolfram’s extrapolations from his work (he believes his findings are relevant not only for hard sciences but also for social sciences) and the unusual interest it has engendered outside the academy. It’s not just Wolfram’s conclusions that are unexpected; it’s also his methods. And such can be said for Stephen Wolfram himself, as a scientist and as a manager.
Indeed, the practices Wolfram the man uses to lead Wolfram the company are anything but conventional. He’s best known for his technical discoveries, but his most ingenious invention may turn out to be a successful company that he built around his own idiosyncrasies: his decisions about where to work, when to work, how to work—in a nutshell, his insistence on building a corporate culture that behaves a lot like he does.
The practices Wolfram the man uses to lead Wolfram the company are anything but conventional.
Conventional wisdom about leadership says that you shouldn’t act like you’re the smartest guy in the room. But how can you do this successfully if you are consistently the smartest guy in the room? Wolfram received his Ph.D. in theoretical physics from the California Institute of Technology when he was 20, two years before he became, at the time, the youngest-ever recipient of a MacArthur Fellowship (in 1981).
Actually, working with Wolfram doesn’t so much entail dealing with someone who’s the smartest guy in the room as it does dealing with someone who’s the smartest guy on the phone. Wolfram Research is based in Champaign, Ill., but its staff members are distributed around the globe, and Wolfram runs the company from his home north of Boston, with only infrequent visits to the headquarters.
Wolfram is a fascinating presence on the conference calls he uses many times a day to supervise the business of Wolfram Research. The five such calls that I sat in on included some low-key project kickoff calls in which Wolfram indulged his curiosity, speculated on what was possible, and paused to see who could add to his insights. There were also middle-of-the-project calls in which the messiness of software development was apparent and Wolfram wasn’t shy about creating a sense of urgency to get past it. During one call, he dismissed a key feature of a program (“That’s amusing, but that’s not top of mind”); in another, he dismissed someone who advocated an approach he wasn’t sold on (“No, let me explain what’s relevant here”). That blunt delivery can cut both ways: A subordinate had no trouble asking his boss, quite directly, “Do you understand what I just said?” But it’s always clear who gets to decide. And it wasn’t surprising to hear that the more Wolfram talks, the more technical the conversation gets, even if the call is ostensibly about marketing.
“Listening to those calls must have been a curious spectator sport,” Wolfram said a few weeks later during an early-evening meeting in a conference room at his home. “I’ve spent more than a quarter of a century finding people who I think are really smart. Maybe I have more experience than they have overall, but in particular areas I certainly hope that they’re smarter than me. The only thing that goes wrong sometimes is when there’s somebody who isn’t really getting it and they’re wasting a bunch of time for the 10 other people who are in the meeting.”
“Getting” what the boss wants is crucial for success at Wolfram Research, a company where Wolfram is the founder, namesake, and primary inventor. “My approach, which maybe not everybody agrees with, is that I’m very straightforward. I just tell people what I think, good or bad. I think people ultimately appreciate that because sometimes they’ve done something that wasn’t very good and I tell them that it’s not very good. Then they redo it or whatever, and it’s actually good.”
It’s not just his technical expertise that makes Wolfram the smartest guy on the phone, according to Samer Diab, the chief operating officer of Wolfram Solutions, the company’s consulting operation. “Someone like me is capable of managing 20 to 25 people,” he says. “The great executives I’ve worked with in the past max out around 150 people [about whom] they know to a detailed level what is happening. More than that and they need a management structure to keep up. Stephen is different. He has reached the ability to manage 600 people reasonably closely even if they’re not direct reports, and we haven’t found his maximum capacity yet. In five to 15 minutes, Stephen can ingest what’s happening with a group and then rapidly move to conclusions and offer advice and direction.”
Diab thinks his working relationship with Wolfram is different from the relationships Wolfram has with those in technology-oriented roles. “He does not inject himself into my operation like he does technology and development,” Diab says. “I talk with him to get the advantage of his brainpower. I tell him the problems our customers are facing and we discuss what groundbreaking ideas can we offer them. I absorb like a sponge while his brain does what his brain does. Then I go off and run my organization.”
Building on a Founder’s Vision
Not everyone can excel in an environment where an idiosyncratic leader is playing a short and a long game simultaneously. Wolfram has a portfolio of R&D projects that range from one (Wolfram|Alpha) that refreshes weekly to one that won’t have an impact, if it ever does, for a generation—as the theory of A New Kind of Science succeeds, fails, or is amended in the marketplace of science ideas.
And yet all those projects are meant to be complementary. Wolfram said that he started Wolfram Research in 1986 to “build tools I knew I would need.” When you look at his three major public achievements prior to the just-announced Wolfram Language—Mathematica, A New Kind of Science, and Wolfram|Alpha—you can see how the different products are enabled by, and improve, one another:
•Wolfram created Mathematica to build a foundation for A New Kind of Science research. It became a successful business on its own, to be sure, but Wolfram had a different motive for constructing the program. And one of his primary means of evangelizing for his New Kind of Science ideas is a summer school in which students test and expand the theory using Mathematica.
•Wolfram|Alpha is a product that its creator sees as having been made possible by A New Kind of Science theory, employing that system’s ideas about combining and retrieving information to create the only search engine rich enough to complement Google’s (although Wolfram detests it when someone refers to Wolfram|Alpha as a mere “search engine”).
•Recent versions of Mathematica integrate Wolfram|Alpha so people can run real-time Alpha searches from inside Mathematica-created programs. It works the other way around as well—snippets of Mathematica code can be used as input in Alpha.
Each success is built on and adds value to previous ones, such that each now feels like a building block in Wolfram’s larger and still-evolving vision. And building blocks, of course, are made to be combined. Little more than a week after he announced the Wolfram Language, Wolfram returned to his blog to reveal that the language, as well as Mathematica, would be bundled with the system software for the Raspberry Pi—the device du jour for young hardware hackers. (He hopes this will get his ideas into the hands of the next generation of tech geniuses.) It’s still unclear the specific ways in which the new Wolfram Language differs from the existing Mathematica language; we’ll see when it’s released and inspected.
Most recently, Wolfram has launched a connected devices project with the aim not only of collecting knowledge about a plethora of smartphones and tablets, but also of serving as the platform on which the data in those devices can be shared and analyzed—by programs like Alpha.
Since Wolfram’s various initiatives are so closely linked, it’s important for the company to find and develop people who can hold the different parts of the founder’s vision in their head simultaneously. “Getting the right talent is important,” Wolfram acknowledges. “But, sadly, these days I don’t do that much frontline interviewing. When I did more of that, we’d talk to the person for a while about all kinds of things and we could tell whether they were capable of expressing themselves, or whether they were bullshitting. We’ve found that as long as people can say what they think and be straightforward and so on, it usually works.”
Perhaps as the company continues to grow, Samer Diab’s depiction of Wolfram as an all-knowing leader intimately connected to all his people and their work will get more complicated, especially since Wolfram is no longer an active part of every hire at his company. And these days, a new Wolfram employee’s development may take a while, regardless of who shepherds that person through the hiring process. “Sometimes, managers tell me that a person is good but not ready to be in a meeting with me yet,” Wolfram says. “Many times, in a meeting, I’ll ask who knows about something. If the person who’s supposed to know about it doesn’t actually know anything about it, it usually turns out badly.”
What kind of person can thrive in such an environment? Cliff Hastings started at Wolfram Research 16 years ago, doing tech support and driving something called the MathMobile, which he brought to universities and colleges so he could demo Mathematica. He’s now the company’s director of sales and strategic initiatives. Hastings identifies the shared characteristics of people who have succeeded there for a long time: They’re unequivocally Type A. They know Wolfram’s products and technology backward and sideways. And when hiring, he says, “we look for a person more than someone who can fill a role. A special person will move beyond the job they were hired for, and a passion for the company turns into passion for a job you helped create. Many of the people we bring in [are hired] because they’ve expressed enthusiasm for one part of what we do. Then they come in and get even more excited about something else. It’s all connected, though. You’d figure that someone excited about Mathematica would also be excited about Alpha.”
Wolfram agrees that hiring the right people can lead to good surprises—even if it sometimes takes a long time for those surprises to materialize. “One of the more bizarre things that’s happened in the history of our company is this thing where I’ll hire talented people and then I’ll realize I’m not really quite sure what this person’s going to do for us. Sometimes they’ll be floating around for several years. Well, there are a few things they can do here and there, but there’s no really strategic thing they can do. Then I’ll suddenly realize, gosh, there’s this new project we’re doing, and this person is the ideal person to be a key figure in that. There are two dynamics. First, you have talented people; then you figure out, as part of the role of management, how to connect these talented people with a project that actually needs to be done.”
Identifying connections is clearly a theme with Wolfram—between products, between people and products, and between people, products, and the whole of the company. “The number one thing I probably contribute is making connections to other things,” he says. “As a CEO, I get different people in different parts of our company to learn about what’s happening in other parts of the company. It’s somewhat successful, but ultimately I’m usually the one who has to tell people to make this or that connection.”
Staying Connected Remotely
Wolfram Research would be a different company if its leader worked regularly at the main office in Champaign. (Wolfram was formerly a professor at the University of Illinois.) “I did live there for the first few years of the company,” he says. “But I like the concept of people being able to live wherever they want. The whole company is very distributed. I think it’s rather healthy that people are not all on top of each other all the time.”
Wolfram travels from the Boston area to Champaign at least three times a year and some random additional times, but not for very long. “I was there last week and I was only there for two days. I was going nuts, because I said there isn’t enough for me to do here. I can sit in my office and do the same readings I would do anywhere else. When I lived there more than 20-something years ago, there were a couple of issues. One was that people would just wander in and say, ‘Oh, by the way, can I talk to you about this?’ And it turned out they didn’t need to talk to me about that. It would have been much better if it was in an email. Then I can process it in an organized way. If they want to have a meeting, they schedule a meeting. I feel vastly more efficient when I’m working remotely than I do when I’m on site.”
What does feel efficient to Wolfram is his rather unusual working regimen. “I have a pretty precise schedule,” he says. He works from 11 a.m. to 6 p.m., breaks until 8:30 p.m., and then works until 2:30 a.m. It’s a 13-hour workday, and he says it rarely varies. (Our two-hour meeting at his home took place during his usual break time.)
That long day with a break in the middle makes it easier for Wolfram to communicate with executives across many different time zones. Luc Barthelet is executive director of Wolfram|Alpha, and he is based in California. “Stephen and I spend several hours a week on the phone,” he says. “We have a one-to-one every Monday evening. Well, it’s evening for me. It’s the middle of the night for him. When I’m going to talk to him, I have to tighten my seat belt. Stephen spends 80 percent of his awake time in conversations. That’s how he thinks. He needs to speak to someone to get ideas. He needs to speak. We don’t have meetings where he just listens. He doesn’t just absorb.”
The remote nature of work for many in Wolfram Research suits Barthelet fine. “I feel like I have an extra day of the week because I’m not commuting,” he says. “I don’t know why [working from home is] not more widely used. It offers dramatically higher productivity.”
Optimizing for Eureka
Wolfram’s audacity is both one of his greatest selling points and one of the most commented-upon aspects of his public image. In an 8,000-word review of A New Kind of Science, Ray Kurzweil, no stranger to hubristic enterprises himself (he’s working on ways to live forever), notes, “I find Wolfram’s enthusiasm for his own ideas refreshing,” even if he concludes that “Wolfram’s sweeping and ambitious treatise paints a compelling but ultimately overstated and incomplete picture.” For his part, Wolfram has been refining his theory and encouraging research into its ramifications. Indeed, Wolfram considers the controversy over his theory a leading indicator of eventual success. He says, “The best predictor of a paradigm shift’s success [over the] long term is how upset it makes people.”
Although projects like A New Kind of Science and creating a new kind of computer language may seem like swing-for-the-fences projects, Wolfram says he doesn’t view them that way. “I don’t see myself as particularly ambitious. I’ve been lucky enough to be able to do some things that have been reasonably successful. Once you’ve done that, you realize, ‘This is fun, I might as well do more of this.’ For me, with my particular psychology, I’m thinking, ‘I’m doing something that nobody else is going to do.’”
Wolfram may seem disingenuous, declining to characterize himself as ambitious, trying to seem cordially modest to an interviewer. Yes, he’s enormously confident of what he’s accomplished and he’s delighted by the work he gets to do, but he sees himself as only “reasonably successful.” In his mind, he’s just getting started. He’s still hungry. And, perhaps more than anything else, Wolfram wants to be right: about his business, about what his products can do, about the new kind of science he believes he has discovered. Chris Anderson, who curates the TED conference and has hosted Wolfram as a speaker, gets close to the core of how Wolfram’s advocates see him when he says, “Just because Wolfram has an ego doesn’t mean he’s wrong.”
Wolfram says that keeping Wolfram Research as a privately held firm is the best decision he ever made as CEO. At a technology conference in Boston in 2010, Wolfram graded his company this way: “I’d give us an A+ in technology and R&D, but maybe a B or C in business.” When reminded of this, Wolfram says, “It’s a question of what one is trying to optimize. For me, I would say that probably since that time we have improved the business side of the company. Because we’re not public, we don’t make as much money as we might, but I’m the one who loses the most [because of that], and it’s my decision. I do feel an obligation to my employees to give them an environment where enough money is being made that they can do well. But you can optimize things to maximize the amount of money that the company makes, and we definitely have not done that.”
What Wolfram has done, though, is build a company focused on identifying big problems and going at them without restraint. “It’s been a very positive thing here, being able to motivate people to work on what seemed like impossible projects. In many situations, you get in some meeting and somebody will say, ‘There’s 30 years’ worth of literature about this and it’s still an unsolved problem.’ At that point, the group might say, ‘Forget it, that’s just hopeless.’ At least we’ve managed to develop a culture where people say, ‘Great, let’s try and solve this.’ The thing that’s really interesting is how often one can.”
Reprint No. 00256