Not titled

Second, we must design for discovery rather than prediction, creating systems (from personal routines to whole cities) that can quickly detect and amplify what works rather than trying to foresee it in advance.