Fruchterman came out of Caltech in 1980 with bachelor’s and master’s degrees in electrical engineering and applied physics (and a private pilot’s license, too). He went on to Stanford to pursue a Ph.D., and like most techies there, was bitten by the entrepreneurship bug. At a series of talks by entrepreneurs, he met Gary Hudson, whose GCH Inc. was building a rocket and intended to compete with NASA’s satellite-launching business.

After his talk, Hudson invited Fruchterman to have dinner. “Gary asked me who my favorite science fiction author was,” Fruchterman recalls. “I told him Poul Anderson. He said, ‘You’re hired.’”

Fruchterman left Stanford and became GCH’s chief electrical engineer. He was 21 years old. His boss there, now one of his closest friends, was David Ross, then a 30-year-old physicist. Fruchterman took charge of designing the rocket’s telemetry, remote control system, and self-destruct command system. In August 1981, GCH trucked its first rocket to Texas for launch.

The rocket exploded on the launchpad. Fruchterman and Ross made sure the fire crews put out the flames and then joined a few other GCH employees who commandeered the company plane and flew it to the Bahamas for a week on the beach. By the end of the week, Fruchterman and Ross had decided to start their own rocket company.

Unfortunately for them, they never got the $300 million in funding they had sought. But while trying to get it, they met Eric Hannah, then a microprocessor designer with Hewlett-Packard. He had the start-up bug as well. He wanted to build an optical character recognition system on a single chip, something that could read anything. But he didn’t have an application in mind.

Fruchterman had a great idea, one he’d been mulling over for years: a reading machine for the blind.

Fruchterman, Ross, and Hannah joined together in 1982 to form what they eventually called Calera Recognition Systems. The firm would make a new kind of optical character recognition device, one that could recognize any font without being painstakingly “trained” beforehand by its user.

Research for their business plan turned up the obvious but disappointing fact that there was no money in making reading machines for the blind; rather, the earning potential was in devices that scanned legal documents, insurance forms, and tax returns. Reluctantly, Fruchterman went along with the shift in the target market.

They raised $25 million, hired 110 people, and by the mid-1980s, they were selling font-independent character-recognition systems to the tune of $10 million a year. And Fruchterman was bored.

So in 1986, he and Ross started a skunk works within Calera to finally build that reading machine for the blind. In 1987, they showed a prototype to their investors, and they argued that although the product might bring in only $1 million a year, it would be good for public relations and employee morale.

The investors said no. Fruchterman and Ross left Calera in 1989.

They obtained the rights to start two companies based on the Calera technology: one, RAF Technology, in Redmond, Wash., to develop custom products for large government customers; the other, Arkenstone, to do the reading machine. In exchange, they had to agree not to hire any Calera employees or compete with Calera for a year.

Fruchterman spent a year dealing with the legal and logistical minutiae of establishing the new companies. He set up Arkenstone as a nonprofit organization, reasoning that at his projected $1 million a year in sales, he couldn’t do better than break even, and there would be public relations and tax advantages to being a nonprofit.

For a while, both Fruchterman and Ross worked with both companies. In 1995, they split them. Ross took RAF, which today supplies optical character recognition technology to the U.S. Postal Service and to other businesses that process mail and forms. It also supplies authentication technology to the U.S. Department of the Treasury. Fruchterman took Arkenstone, which then had $5 million of annual sales from reading machines, at about $1500 each.

It was fun. For a while. In the late 1990s, Fruchterman says, “I’d been running a $5-million-a-year business for 10 years. I wanted to do more.”

Around that time, he and Ross would meet regularly for long hikes through the foothills above Silicon Valley, while lengthily discussing a broad range of subjects. On one of these excursions they got to talking about various human rights outrages—including a massacre of 600 people near the Sumpul River in El Salvador in 1980 and the bloodbath in Tiananmen Square in 1989. They speculated about futuristic technologies that could protect protestors or at least alert the world if these people were killed or abused.

“From then on, we knew we wanted to do something in the human rights field,” Fruchterman said. And with the boom starting to rumble, “we suddenly had all these dot-com billionaires who we thought might give us money.”

Meanwhile, Freedom Scientific, a for-profit company based in St. Petersburg, Fla., that makes products for the blind and vision-impaired, made an offer to buy Arkenstone. The timing couldn’t have been better.

In 2000, Fruchterman formed the nonprofit company Benetech, funded with the $3 million he got for Arkenstone from Freedom Scientific.