The discussion centers on the theoretical maximum state changes per second for optical computing transistors, with estimates ranging from 200 billion to potentially 500 billion state changes. Participants debate the feasibility of achieving such speeds, with some suggesting that light could theoretically allow for up to 4 quadrillion state changes per second based on the number of atoms light interacts with. However, others argue that practical limitations and the nature of photons prevent such high rates from being realized in current technology. The conversation highlights the complexity of optical computing and the challenges in translating theoretical calculations into practical applications. Ultimately, the consensus leans towards a more realistic expectation of 10-20 GHz for practical optical computing applications.