Learning from the CMB with new data and differentiable theory

Abstract: I will discuss three strategies for making new discoveries from the Cosmic Microwave Background: take more data, improve our analysis, and accelerate theory development. I will share my excitement about the next generation of CMB experiments, and give an update on the progress of the Simons Observatory. I will also describe the development of a new suite of mock galaxy simulations to enable joint analyses of the CMB with large-scale structure. Finally, I will provide a first look at Bolt, the first differentiable Einstein-Boltzmann solver. Bolt accelerates the exploration of the space of theories, by computing theory predictions like CAMB and CLASS along with exact gradients of observables with respect to input parameters.