Why Don’t More CS PhD Breakthroughs Turn into Companies?
Silicon Valley is acclaimed as an innovation hub in the public eye, but dramatic technological breakthroughs occur in the halls of university research labs all the time. PhDs have years of deep knowledge in narrow domains that often produce discoveries that propel tech advancements even further.
Academics fathered some of the most important innovations that lead to the commercialization of computing. Before mathematician Alan Turing published his theory of the ‘Universal Turing Machine’ at Cambridge University in 1936, there was no concept of storing a program in a computer.
The Universal Turing Machine is the blueprint on which the modern computer is based today. At a time when different machines were dedicated to accomplishing different tasks, the idea of a universal computer that carried out any task on one machine was remarkable.
But it was only hypothetical and remained as such until the first prototype “Pilot ACE” (pictured above) was created in 1950 by other researchers who took over Turing’s ideas. Even still, it wasn’t commercially sold until the company English Electric created a version of the Pilot ACE and sold it as the DUECE in 1955.
Sure, there are some exceptions of research ideas that prospered into successful business ventures. The most famous example of this is Google by computer science PhDs Larry Page and Sergey Brin. Before Google, web pages were ranked in order of websites with the highest number of key terms. Page applied the concept of building authority through citations (or backlinks) and theorized a smarter system of organization dubbed “BackRub.”
He and Brin created an algorithm that rewards links that come from credible sources. Backed by support from their colleagues at Stanford, they launched the first version of Google on the university website: www.google.stanford.edu. At one point, the duo even crashed the university’s Internet connection!
“We’re lucky there were a lot of forward-looking people at Stanford,” Page tells WIRED. “They didn’t hassle us too much about the resources we were using.”
They received modest offers for the technology from existing companies in the space, like Yahoo!, Infoseek, Lycos and AltaVista. Encouraged by their colleagues, they turned them down and took it upon themselves to take the plunge as entrepreneurs. By and large, however, most breakthroughs are very slow to materialize from the lab into the industry. Today thousands of research labs around the world are producing new techniques and even products that could potentially be valuable to the industry and economy. But many PhD papers remain, well, papers.
Consider a recent example at MIT, one of the world’s most prestigious computer science doctorate departments. In 2010, a research team designed an intelligent wheelchair with voice-commandable robotics. It uses machine learning to automatically learn the layout of any given room and carries out command by voice. It could help thousands of disabled people who suffer brain injuries who can still talk. Five years later, there’s still no sign of any such product to hit the market officially yet.
Here’s another smart wheelchair conceptualized at Worcester Polytechnic Institute, reported in 2014. This one allows you to control the wheelchair’s motor by raising your eyebrows. It lives in the university’s robotics workshop for now, although the inventor Taskir Padir says he’s aiming to push out a commercial version of the navigation system in the next few years. Why aren’t we seeing such technological breakthroughs hitting the market more frequently today?
Few Universities Offer Entrepreneurial Support
Elite universities, like MIT, Stanford and Harvard, may have a strong network and support system to help their researchers progress select breakthroughs into the market. But most universities leave it up to PhDs to seek out investors, business-savvy leaders and all other moving pieces that are necessary to launch a successful business.
Stony Brook University Distinguished Teaching Professor Steven Skiena is one academic who successfully turned his expertise in Big Data sentiment analysis into a company. He says, it’s all about who you know.
“The key event in founding General Sentiment was when I met an experienced entrepreneur (Mark Fasciano) with the experience and inclination to make the venture happen,” he says.
Building a team and getting the right investors on board requires time that most PhDs just don’t have. “If universities provided support and encouraged not just to patent inventions, but actually take them to market, we would see more startups,” says Yevgen Borodin, research assistant professor at Stony Brook University.
It takes the right leader to launch and turn it into something economically productive. More investors and companies should work to bring research ideas to life. Rishabh Jain, MIT PhD, offers up 3 examples of Venture Capitalist firms that fund university breakthroughs, according to Quora:
Some Papers Aren’t Practical or Satisfy a Market Need
There’s a stark difference in objectives between businesses and research facilities: “Most companies are not really based on ideas, but instead, recognizing and satisfying under-served market needs,” Dr. Skiena says. “They are different beasts: both important, but different.”
Timing is a critical factor here. Publishing new ideas is generally a slower process compared to the startup world. Without investors with an eagle eye on revenue and profits, scientists have the luxury of working on ideas for the sake of knowledge. This freedom often results in breakthroughs that are important to technology long-term rather than filling an immediate market need.
Achievements in newest domains, like artificial intelligence, are a culmination of several different ideas over several years. Take computer vision researchers at MIT, for instance, who recently figured out an algorithm that removes the reflection that appears when you take a photo through a window. While that’s really cool, the discovery is only effective on double-paned windows. With such limitations in place, the breakthrough alone may not permeate the photography lens market anytime soon, but it’ll be pivotal for other robotics creators in improving robot vision of the future.
On the other hand, Dr. Borodin creatively turned his product into something more immediately practical for the wider public. His research aimed to help blind people interact with computers. Determined not to let his research be confined to ideas on paper, he founded Charmtech Labs LCC with a few other colleagues and marketed the product himself.
“What helped us succeed in the market was the decision not to develop technology for blind people alone, but, instead, make it universally accessible by everyone,” he says. Now, his product Capti Narrator is marketed as a productivity tool for anyone to turn documents or web pages into audio.
More PhDs Should Consider Pushing Breakthroughs to the Market
Even after PhDs spend years devoted to unique tech discoveries, some often have a hard time getting jobs that are directly related to their research. The highest rank of a PhD is getting tenure, but statistically there are 10 PhD graduates in computer science for every 1 tenure position.
The most recent survey by the Computer Research Association finds that in 2014 the number of doctoral degrees produced declined by 2.6% from 1,991 to 1,940. Here’s a graph describing the trends in PhD production over the past few years. As you can see, the number of PhDs have generally stalled, compared to its initial spike between 2003 and 2007, when the number of PhDs just about doubled:
Taulbee Survey finds that 57.5% of PhD graduates went on to work in the industry in 2014, up from 55.5% in 2013.
Granted, in academia, one key objective of discovery research is purely for the advancement of knowledge. So, some researchers have no intention of creating breakthroughs for commercialization.
But academics who are considering leaving academia should start thinking broadly about how to apply their breakthroughs in the industry, whether by launching their own startup or accommodating for a larger market. Likewise, universities should provide a better springboard for connecting researchers with the right resources to move their ideas forward. There’s a lot of room for improvement when it comes to reducing the lag time between PhD breakthroughs and commercial products that could help accelerate tech advancements for all.
What other reasons have you found in explaining why more PhD breakthroughs don’t turn into companies?