SUNNYVALE, Calif. — Lee Redden, 26, a Ph.D. student in engineering at Stanford, recently decided to shelve his education and help found a start-up company. His skills lie in a couple of red-hot niches of artificial intelligence, computer vision and machine learning. Yet he is not applying his talents to Internet search, online commerce or intelligence surveillance.
Mr. Redden’s ambitions are further afield — in farm fields, actually. His company, Blue River Technology, is developing a robotic weed killer for organic farms, which shun chemical pesticides. The new venture, he said, is “a great way to bring this technology to agriculture.”
The start-up here points to the latest stage of evolution in Silicon Valley, the world’s epicenter of innovation. Over the years, the region has shown an unmatched economic dexterity in jumping from one industry of opportunity to another, from military electronics to silicon wafers to personal computers to the Internet.
But the business of the Valley today is less about focusing on a particular industry than it is about a continuous process of innovation with technology, across a widening swath of fields. The trend reflects the steady march of that most protean of technologies — computing — as it makes further inroads into every scientific discipline and industry. Clean technology, bioengineering, medical diagnostics, preventive health care, transportation and even agriculture are part of the mix these days for the Valley’s technologists and entrepreneurs.
The pace of discovery has quickened, not only for technologies but also for the process of finding out what companies will succeed.
“What’s different in the Valley is that we’ve found a quasi-scientific method for reinventing businesses and industries, not just products,” said Randy Komisar, a partner in a leading venture capital firm, Kleiner Perkins Caufield & Byers, and a lecturer on entrepreneurship at Stanford University. “The approach is much more systematic than it was several years ago.”
The newer model for starting businesses relies on hypothesis, experiment and testing in the marketplace, from the day a company is founded. That is a sharp break with the traditional approach of drawing up a business plan, setting financial targets, building a finished product and then rolling out the business and hoping to succeed. It was time-consuming and costly.
The preferred formula today is often called the “lean start-up.” Its foremost proponents include Eric Ries, an engineer, entrepreneur and author who coined the term and is now an entrepreneur in residence at the Harvard Business School, and Steven Blank, a serial entrepreneur, author and lecturer at Stanford.
The approach emphasizes quickly developing “minimum viable products,” low-cost versions that are shown to customers for reaction, and then improved. Flexibility is the other hallmark. Test business models and ideas, and ruthlessly cull failures and move on to Plan B, Plan C, Plan D and so on — “pivoting,” as the process is known.
The National Science Foundation is betting on the new model to improve the rate of commercialization of the university research it finances. In October, the foundation announced the first series of grants for what it calls the N.S.F. Innovation Corps. The 21 three-member teams selected from across the country will receive $50,000 each for six months to test whether their inventions are marketable. It begins with a swing through Stanford and courses taught by Mr. Blank and others, followed by online classes and mentoring. Each team is expected to constantly test its ideas and products with customers, to experiment again and again, adhering to the lean start-up formula.
“It’s all about how to apply the scientific method to market-opportunity identification,” said Errol B. Arkilic, a program manager at the foundation. “And that is exactly why this method is the one the N.S.F. selected.”
Mr. Arkilic, who worked for seven years as an engineer at start-ups in the Valley before joining the government, says that the foundation plans to award 15 or 20 Innovation Corps grants every quarter. “We can’t replicate Silicon Valley elsewhere,” he said. “But we need to figure out a way to take some of the best practices in Silicon Valley and deploy them elsewhere.”
The Valley’s lean start-up model is influencing mainstream business education. At business schools, courses on entrepreneurship and entrepreneurial management have been around for years. But this spring, the 900 first-year students at the Harvard Business School must start a business as a required course. In teams of six students each, they will be given $3,000 and told to create a start-up that pulls in revenue by the end of the semester, explained Thomas R. Eisenmann, a professor who will oversee the program.
Dr. Eisenmann is also a leader of a Silicon Valley immersion program that takes dozens of students there each year for a firsthand look at not only the practices but also the culture of the entrepreneurial hothouse. “Everybody is involved in a start-up,” Dr. Eisenmann observed. “It’s assumed to be normal behavior, the cool thing to do.”
The start-up culture, to be sure, owes much to the Valley’s history and the enduring influence of its defining personalities, like Frederick E. Terman. A longtime engineering professor at Stanford who was provost in the 1950s and 1960s, Dr. Terman encouraged his star students to put their ideas to work by starting their own companies. Among them were Bill Hewlett and David Packard, who founded Hewlett-Packard in 1939. Dr. Terman often invested his own money in such start-ups.
“His message was that starting a company was every bit as important as getting your Ph.D.,” said Mr. Blank, who lectures at Stanford. “That was heresy in academia.”
Most start-ups fail. More than perhaps anyplace else, the Valley’s investors and technologists tend to view failure with a certain scientific objectivity, as if setbacks are heuristic tools leading to further research and discovery.
“Over 70 years, this Valley has developed a culture that does not personalize failure,” said Mr. Komisar, of Kleiner Perkins. “If you’re not corrupt, stupid or lazy, we see failure as learning — learn from it, and reapply it.”
That risk-taking and perpetual learning culture is now being applied far beyond computer hardware and software. There are high-profile ventures like Nest Labs, founded by Tony Fadell, a former Apple executive, which has hired more than 100 engineers from Apple, Google, Microsoft and other high-tech companies, and it is backed by a pack of venture capital firms. Its product, introduced in late October, is a reinvention of the thermostat, combining sensors, machine learning and Web technology in a clever, energy-saving device.
Yet the Valley is filled with other, less-prominent start-ups in new fields, including Blue River Technology. Mr. Redden met his fellow founder and the start-up’s chief executive, Jorge Heraud, 41, in Mr. Blank’s class.
Mr. Heraud, an engineer, had decided to return to Stanford to refresh his career, after working for years as a manager for an agriculture equipment company.
He met up with Mr. Redden, and they jointly recognized the opportunity for applying computer vision and machine learning to agriculture. They explored other ideas, but after talking to potential customers, they settled on killing weeds for organic farming — a fast-growing industry that cannot use pesticides and struggles with labor costs.
In their office in Sunnyvale, they are assembling the machine and training the computer-vision software to distinguish lettuce plants from weeds. Initially, they experimented with laser blasts to kill the weeds, but that proved too costly. Now they plan to use superheated oil (organic, of course).
Their contraption, pulled behind a tractor, must identify weeds and kill them within 200 milliseconds, a formidable but attainable challenge, Mr. Redden says.
If applying computer vision to agriculture is so promising, why are the big farm equipment companies not well ahead?
“This isn’t on their radar screen,” Mr. Heraud replied. “They aren’t putting their heads into computer vision. That’s a Silicon Valley technology.”