Find in Library

The cumulative development of culture has proven difficult to study in the laboratory. Here, the authors examine entries to a series of large programming contests to show that successful entries are usually ‘tweaks’ of existing solutions, but occasional ‘leaps’ can bring larger benefits.