The first AGN was identified as a variable source over ninety five years ago and their extragalactic nature was determined over sixty years ago. Today we have catalogs of millions of them and the next generation of sky surveys will detect tens of millions but their inherent variability across all accessible wavelengths and timescales remains a profound puzzle. We have no strong indication what the underlying physical processes are, we cannot infer basic physical values from time series, and we cannot forecast future activity with any great degree of accuracy. Attempts over the past decade to address this through increasingly complex statistical and machine learning models have not met with stunning success and our best simulations still lack sufficient resolution and complexity. However, a recent shift of focus to studying the broader environment of supermassive black holes and on interactions between their different elements is proving more fruitful. In this talk, I will explore the latest observational evidence into AGN variability, particularly focusing on extreme events and what insights they offering us into accretion processes and the surrounding circumnuclear environment. Some of these are particularly important as false positives in periodicity searches for supermassive black hole binaries.I will also discuss the potential for upcoming observational facilities and advanced techniques to help further our knowledge of AGN variability.