The evolutionary model of competitive selection is hard to translate in healthcare where current culture, incentives and policies often lead to a failure to check if something works and act on the results. This is particularly problematic in areas of high uncertainty (and corresponding high risk of failure for any proposed strategy), like the care of people with complex needs. We look to the software sector as an example of a human ecosystem experiencing an explosion of diversity that facilitates participation of people from varied backgrounds and has strong selection processes and approaches to manage uncertainty. Key lessons from this sector include facilitating failure through rapid tests with ready alternatives, support for people (not just ideas) so they can try different approaches and a system-level portfolio investment to account for high likelihood of failure of any given project. A successful ecosystem in healthcare would not only select proven strategies, but promote collaboration among innovators so that there is cumulative system learning as opposed to personal empire building. Given the major fiscal, social and demographic challenges on the horizon, failure to search for novel solutions is a bigger risk than trying new things that might not work.