During most of it's history, software design has basically tackled the problem of creating tools that enable single users to perform a set of tasks in order to achieve a specific, predetermined goal. Though potentially large numbers of users could be using the same system, the process of getting to a result is reached through tasks performed by individual users. The advent of crowdsourcing (using the broad definition of the word), has so far mostly split large tasks in smaller pieces that can be then distributed among a big number of people, something also known as micro-tasking. Succesful examples like Amazon Mecahical Turk abound, but in almost all of the cases, the nature of the goal is directly de-composed into the tasks, meaning that the whole will never be greater than the sum of the parts. I suggest that by taking an interdisciplinary approach and learning from game design, system thinking, genetic programming, market economy and behavioural sciences a new breed of systems could emerge that would tap in the power of the crowd in interesting new ways. By designing systems built on rule-based models and allowing for emergent behaviours, the exploration of much more complex and less deterministic problems could be possible, including many of the world most pressing challenges. I propose to analyze a few preliminary examples and suggest ideas for further exploration.