Major changes in the world have led to what some scholars are calling ‘post-normal times’, an age characterised by uncertainty, unpredictability, rapid change, complexity, chaos and contradiction. Some have argued that, as a result of these developments, major reform is required in education. The reforms argued for are widely known as ‘future-focused’ education. In the past, education reform efforts have often encountered unanticipated complexity which has derailed or diluted their implementation. This thesis argues that past reform efforts have been limited by their assumption of a reductionist paradigm. It describes a study designed to test the usefulness of complexity thinking for understanding and implementing the kinds of changes needed in ‘post-normal’ education. The project involved designing a ‘safe-to-fail experiment’, in the form of a MOOC, to investigate whether the system could be ‘nudged’ to produce conditions under which future-focused change can emerge. The effects of this experiment were investigated via a complexity-informed case study of some of the MOOC participants. The study found that the system was influenced by this ‘safe-to-fail experiment’. However, deeper analysis revealed a more nuanced system of negative feedback loops acting to constrain the possibilities for system-wide, future-focused change. This study concludes by suggesting that complexity-oriented research approaches are probably most useful, at this point in time, for identifying and exploring the nature of the system’s homeostasis-maintaining feedback loops. Such attempts to ‘see the system’ are likely to suggest fruitful sites for further experimentation.