The vast majority of the calories (70%) we consume on a daily basis are derived from grains (wheat, rice, maize). However, these crops are input intensive (capital, energy, and physical labor) and very vulnerable to shortages/disruption of those same inputs (mostly because they are annual crops that need to be replanted every year). In order to make them appropriate for integration into resilient communities, a new approach is needed: one that allows us to avoid, reduce, or obviate the hard inputs required for grain production through the adoption of a methodology that enables rapid cycles of decentralized innovation.
One potential approach to grains production that meets this hurdle is being pioneered by Jerry Glover at the Land Institute of Kansas (I had the pleasure to meet Jerry yesterday in Aspen). He's working on breeding perennial versions of common grain crops. This provides the benefit of not only increasing the resilience of these plants, the process being used to achieve these perennials is something that would work extremely well within an open source, decentralized tinkering network.