In the 1960s and 1970s, the dynamics of the global climate system were investigated using models that exhibited a high degree of idealization. In his groundbreaking article in Tellus in 1976, Klaus Hasselmann raised the question of whether these models only showed abrupt transitions because they oversimplified the dynamics of the climate system and did not reflect natural climate variability. In this chapter, Hasselmann’s assumption is therefore re-examined. To this end, three examples of multiple equilibria and rapid transitions in the climate system are discussed: Snowball Earth, the Atlantic meridional overturning circulation and the interaction between climate and vegetation. The discussion of these examples begins with predictions from idealized models, followed by evidence of abrupt transitions found in palaeoclimatic proxies that stimulated the search for unstable feedbacks using complex models. Discussion of these examples start with predictions from idealized models, followed by evidence of abrupt transitions found in paleo climate proxies which fostered simulation of unstable feedbacks using complex models. Climate scientists who have worked in Hamburg have made significant contributions to all of these examples.