App-scale science

The 18–24 January 2014 issue of the Economist contained a 14-page survey of tech startups. Ludwig Siegele, the magazine’s online business and finance editor, was the author.

Siegele contends that current-generation startups are different from their first-generation predecessors, such as Amazon, Facebook, Google, and—lest we forget—Pets.com . Whereas launching, say, Google required considerable investment in people and infrastructure, today’s startups can get off the ground more quickly and cheaply by exploiting a range of existing products and services.

The online car service Uber is a prime example of the new style of startup. Launched in San Francisco in 2010, the service connects customers to cars via a smartphone app. Uber’s founders took ingredients that already existed—cars, smartphones, the Google Maps APIs —and cooked them into a new product that filled a need: reliably summoning a ride when you need one.

And it’s not just the use of APIs, Bluetooth chips, cloud computing services, and other preexisting ingredients that characterize the new startups. The way they come into being is different, too. Would-be startup founders join so-called accelerators that provide office space and mentoring in return for a slice of equity. (The most prolific accelerator, Y Combinator of Mountain View, California, takes 7%.)

Of course, if all your startup does is lightly combine existing ingredients to create an app, it’s unlikely to disrupt whole sectors, as Amazon has done to the selling of books and other products. The last app I bought, TapeACall, makes it possible for me to record both sides of conversations on my iPhone. Thanks to the app, I no longer need a separate voice recorder when I interview physicists on the phone. This is convenient, but not transformative.

One of the most interesting aspects of Siegele’s survey concerns the shifting relationship between entrepreneurs and investors. By offering office space, legal advice, and other standard services through accelerators, investors have come to resemble managers while entrepreneurs have come to resemble workers—at least according to consultant Venkatesh Rao, whose writings Siegele quotes.

As I pondered Rao’s observation, I realized that most of the scientific enterprise in the US and elsewhere runs along similar lines. The National Science Foundation and other agencies invest speculatively in researchers who seek to create new knowledge by combining and extending preexisting ingredients, such as lasers, ion traps, and theories. To obtain funding, researchers have to pitch their proposals, just like the founders of startups. Researchers are also obliged to return some of the funding agency’s investment in the form of reports and research papers.

By likening publicly funded science to tech startups, I don’t mean to imply that projects funded by a typical NSF grant are unlikely to be transformative. Nevertheless, the steady reduction in inflation-adjusted funding for science in the US has made funding agencies more conservative.

If that trend continues, the appetite for taking bold risks will shrink—and the US could end up producing only app-scale science.

This essay by Charles Day first appeared on page 72 of the March/April 2014 issue of Computing in Science & Engineering , a bimonthly magazine published jointly by the American Institute of Physics and IEEE Computer Society.