Funding is key to the scientific enterprise, but with budgets stretched and ever more proposals being put forward, the way in which the most forward-thinking projects are approved by funding bodies needs to evolve. The US National Institutes of Health (NIH), for example, has been criticized for playing it safe when handing out its grants and, from its own figures, seems more inclined to fund established researchers than new kids on the block. As the agency is the main funder of US biomedical research, this could prevent or discourage young talent, armed with new ideas, from staying in academia.

At the same time, science is evolving and is providing funding agencies with fresh headaches. Many projects now hinge on interdisciplinary and translational research — and multidisciplinary institutes are springing up around the world. Yet there are too few interdisciplinary-minded reviewers in the United States and elsewhere to handle the proposals for this evolving frontier in research.

The NIH is striving to address some of these issues. In the past year it has altered its grant-review system and is pondering further changes. Its 'Director's pioneer' awards and 'Pathway to Independence' programme are aimed at fledgling scientists. Further changes being considered include a shorter time for application processing, reformed study sections and a 'prebuttal' that would let applicants clarify misconceptions about their application before it goes to review (see http://tinyurl.com/38egx4).

All scientists have a stake in debates on how to determine the most promising scientific ideas. The European Union's Framework programmes are famous for an impenetrable grant-application process that requires complex paperwork and encourages, some would say, unwieldy collaborations. Even so, interdisciplinary science is a big part of the most recent Framework's mission (see Nature 446, 104–105; 2007). Science collaborations have become more complex and, in some cases, bureaucracy has become an impediment to funding avant-garde research easily. Grant review needs to evolve along with the scientific enterprise — otherwise that enterprise will not live up to its promise.