In 1938, two chemists in Nazi Germany fired neutrons at a piece of uranium to discover nuclear fission. Albert Einstein warned President Franklin D. Roosevelt about the emergence of this new force, capable of unprecedented destruction. Roosevelt responded with the Manhattan Project, which beat the Nazis to an atomic weapon and catalyzed a partnership between government and the private sector that led to American technological preeminence for generations.        

Today, two trends threaten that preeminence and our national security. First, the United States and its adversaries are in a Manhattan Project-like race to develop several game-changing technologies, including quantum computing, artificial intelligence and biosynthesis. Second, the United States’ historical lead in basic research and innovation has been fading fast. China is becoming a technological peer. This is a result of the Chinese strategy of massive coordinated investment in designated technological areas.

The stakes are as high as they were during World War II. Artificial intelligence could automate dangerous jobs, eliminate human error from medical emergencies or predict the weather. It could also be used to deploy thousands of autonomous microdrones for violent attacks against which defense would be nearly impossible. Biotechnology is creating ways to manipulate the very building blocks of life that might address pandemics and world hunger or to develop weapons that target specific human genotypes.

Innovation is the product of a complicated mixture of money; inspired, hard-working people; and the environment in which they meet. Many of the ingredients, like unorthodox thinking, tolerance of failure, and flat, flexible organization, are unusual or even anathema to government culture. Government is not naturally configured to “think different” or to “move fast and break things.” That must change.

Most importantly, we do not need to “beat China at its own game” of centrally directed, hierarchical, planned innovation. Instead, we need to do better in the distinctly American directions of openness, flexibility and agility.

Increasing money for basic research is important, but it is far from sufficient. The private sector and other countries are now innovating and undertaking the majority of global R&D. Capturing the brainpower, ideas, methods and products of U.S. entrepreneurs and academics is the key.

This is not a foreign idea to national security leaders. The CIA’s In-Q-Tel venture invests in promising technologies side by side with venture capitalists. The Air Force has partnered with Silicon Valley’s Pivotal Software to pioneer a modern software development unit in Boston. But there are very few programs to rotate outside engineers, academics and data experts into government agencies and departments to solve hard problems. Combine this with the Trump administration’s hard turn against immigration and we are purposely denying ourselves access to the best and the brightest.

Rethinking government policy is particularly crucial in the esoteric area of software development. Software is often considered a product to be acquired. But it is much more. The environment in which it is developed — its speed, agility, responsiveness to the end user and constant, real-time improvement — is essential to almost all other innovation. Marc Andreessen famously noted in 2011 that “software is eating the world”. He meant that every aspect of life, including national defense, has become deeply dependent on software. Government must evolve away from last century’s practices of years-long development cycles for proprietary and cumbersome software to a world in which software is nimble and alive to constant improvement, like your iPhone apps.

Government must also rethink its own structure. Cyber issues and artificial intelligence defy placement into neat jurisdictions. Congress punishes risk in a way that deadens innovation. Woe to the program manager asked to explain failure to a congressional committee, even if that failure clarifies the path to success. And sadly, members of Congress are usually well behind their children and grandchildren in technical know-how.

Finally, winning is important, sometimes critical. But technological races rarely stay won; four years after the Manhattan Project, the Soviets detonated their own atomic bomb. The U.S. must therefore return to leading the world in the establishment of ethics and international norms that guide when, how and why we use these incredible technologies. While we may not always be able to outspend or outman our competitors, we can — and must — do what we’ve always done: lead in the creation of a better and safer world.

Openness, flexibility, speed and leadership are the key. Now is the time for their adoption. Our tradition of technological leadership and the security that follows is at stake.

Rep. Jim Himes is a Democrat representing Connecticut’s 4^th District. He is a member of the House Financial Services and Intelligence committees and chairs the latter’s Strategic Technologies and Advanced Research Subcommittee.