 Hi, my name is Mark. This work investigates the role of temporal coding in everyday hearing using machine learning. Neurons can encode information and the timing of their spikes in addition to their firing rates. The fidelity of spike timing is arguably greatest in the auditory nerve whose action potentials are phase locked to the fine-grained temporal structure of sound. But the role of this exquisite temporal coding in hearing remains controversial. Part of the challenge is that it's extremely difficult to manipulate this phase locked spike timing in a living system, particularly one with complex auditory behavior. We do, however, have detailed computational models of the ear that can faithfully simulate auditory nerve representations of arbitrary sounds. This is a simulated auditory nerve representation that's sound below. It's configured to approximate the fidelity of temporal coding in a healthy human ear. In this work, we ask what might happen if the ear had worse temporal coding? We lowered the phase locking limit from 3,000 to 1,000 Hertz, then to 320 Hertz, and to 50 Hertz, eliminating virtually all phase locking into temporal fine structure. This manipulation decreases the temporal precision of phase locked spike timing. And the question we're interested in is, can a ear without high fidelity temporal coding still support everyday hearing? I'm particularly interested in this question because contemporary cochlear implants discard the temporal fine structure of sound wave forms and a reduced ability to use temporal fine structure is often proposed to underlie speech perception difficulties in hearing loss. To investigate these issues, we optimized contemporary deep learning models to perform real world hearing tasks from simulated auditory nerve representations of natural sounds, asking whether phase lock spike timing is necessary to obtain human-like behavior. We applied this approach to a broad range of tasks, finding that phase locking is necessary to replicate human robustness to sound level and background noise, but degraded phase locking impaired performance more on some tasks than others. Our results link neural coding to real world perception and may help clarify conditions in which prostheses that fail to restore high fidelity temporal coding couldn't principle restore near normal hearing.