As recognised by this call for proposals, we still lack a deep understanding of the laws of Nature. Quantum mechanics is our most accurate model of Nature, but being indeterministic it challenges our notions of cause and effect, the very foundations of the scientific method. However, this fundamental indeterminism opens up a new freedom: the possibility of having non-locality without contradicting relativistic causality. Based on work in quantum information, non-locality is now recognised as one of the basic, if not the most important, aspect of quantum mechanics. However, Popescu and Rohrlich showed that correlations even more non-local than what is allowed by quantum mechanics would still not violate relativity. So why is Nature not more non-local? What is the basic principle of Nature that provides but limits non-locality? We aim to answer these questions, building on a number of very recent exciting results, related to the information processing capabilities of non-local correlations, classical, quantum or stronger-than-quantum. In particular, we will classify proposed 'principles' determine the value of non-local correlations as a resource in communication and computation.