The National Reconnaissance Office, the spy-satellite agency that's so secretive the U.S. government denied for years that it even existed, seems to be catching the entrepreneurial bug that's suddenly infecting the intelligence community.

Early this month, the NRO set up a public Web site to announce a program called the "Director's Innovation Initiative," which will provide $350,000 "seed funding" grants for outside projects that may lead to "revolutionary concepts and ideas."

The goal of this internal venture-capital fund, the NRO posting explains, is to "provide a risk-tolerant environment to invest in cutting-edge technologies and high-payoff concepts" and to "push the boundaries of technology" to improve the NRO's satellite reconnaissance capabilities.

Translation for non-geeks: The spy-in-the-sky spooks want to find new ways to follow terrorists, find their hidden biological and nuclear weapons and track their assaults--whether they come by missile or low-flying drones or truck bombs or suitcases. And the NRO's ability to develop these new technologies, to put it bluntly, is what's going to save Americans' lives in the 21st century.

The NRO's announcement comes a few weeks after press reports about the Central Intelligence Agency's new venture-capital fund. The CIA start-up, known as "In-Q-It," will try to link the agency better with Silicon Valley.

These initiatives signal change for an intelligence community that was in danger of losing America's greatest strategic advantage--its technological edge. The intelligence mandarins realized they needed to reach out to a wider community than the traditional, button-down white-shirt world of classified contractors--and that they needed to take more risks.

"We expect a high failure rate," says Bob Pattishall, director of the NRO's advanced systems and technology directorate, which was created three years ago to push the technology envelope. "If everything worked, it wouldn't be revolutionary," he explains. Pattishall says the agency is "looking for a few home runs" and knows that will be impossible without a lot of strikeouts.

That sort of talk is routine in Silicon Valley, where technology entrepreneurs understand that failure is a requirement of eventual success. Indeed, you can't find a successful entrepreneur in the Valley who hasn't blown it, big time, along the way.

But this kind of risk-taking is rare indeed in Washington--which over the years has ossified into a zero-defect culture, where bureaucrats are constantly looking over their shoulder and worrying about second-guessing from Congress and the media. Any government program that deliberately builds in the need for failure is a step in the right direction.

The NRO is also emulating Silicon Valley by moving at what, for a government agency, is warp speed. The initiative was announced Oct. 4; proposals must be submitted by Nov. 17; the NRO promises to finish evaluating the proposals by Dec. 31 and to make the awards by Jan. 31, 2000. The awards themselves will extend for just nine months, "because we don't want to lock the dollars up," explains Pattishall.

A final laudable feature of the NRO's Innovation Initiative is that it's unclassified--and most of the projects it funds this year will also be unclassified. That's a huge change for an agency that didn't even disclose its existence until 1992. The Innovation effort was actually started three years ago--but within the traditional strictures of NRO secrecy. This is the first year they decided to open it up, with an announcement in Commerce Business Daily and the public Web site. In the first eight days, the Web site received 20,000 hits and more than 10,000 requests for information.

"The NRO is in transition," explains Pattishall. "We haven't been out of the closet for too many years. The goal of this program is to reach as many smart people as we can."

Want to help protect your country from its enemies? Here are some of the areas where the NRO is looking for breakthrough ideas: new ways to "map, plan, understand and execute operations in urban environments," new software such as "intelligent agent-like tools and techniques that integrate acquired information," new ideas about "target phenomenology and related sensor technologies in order to identify target vulnerabilities and new collection opportunities," and new technologies "to detect locate, identify, characterize and track weapons of mass destruction and other advanced weapons systems."

Complicated stuff. But this is the ball game, in terms of our national security. If the NRO and other intelligence agencies can't get the best minds in the country to help them crack complex problems like these, then we're all in trouble.

"Let the sun shine in" is an unlikely motto for the government's most secretive intelligence agency, but it's a welcome one. Here's to a lot of noble strikeouts along the way to a grand slam.