Jets produced from high-energy quarks and gluons are ubiquitous at the Large Hadron Collider (LHC). These objects can be used to study emergent quantum properties of the strong force as well as search for new particles and forces beyond the Standard Model. As a jet can have O(100) particles, analyzing jets is inherently a high-dimensional problem. Therefore, jet physics has been leading the integration and development of modern machine learning tools for high-energy physics. This high dimensionality also is a challenge for classical techniques to account for all quantum effects in the evolution of jet formation. I will start by discussing the exciting new field of precision jet substructure, with the latest results from the ATLAS experiment and interpreted in the context of new theory calculations. This sets the stage for two exciting parallel developments where quantum computers and machine learning may lead to fundamentally new insights. After briefly mentioning the potential of quantum algorithms, I will illustrate the power of deep learning with a new class of algorithms called weak supervision that can learn directly from (unlabeled) data and potentially uncover high dimensional structures hidden from our ordinary three-dimensional view. In this way, machine learning can help us learn something new and fundamental about nature.