Light is a critical medium for quantum technologies, providing the capability both to encode a large capacity of information and to preserve quantum features, such as entanglement and reduced noise, at ambient conditions. Photonics has proven to be an effective means to engineer quantum systems that can perform tasks efficiently in both analog simulation and digital architectures. The challenge is to build a quantum state of sufficient complexity, both in the number of photons and the number of modes, that useful applications can be addressed. I will discuss how optical platforms can provide potential advantage in hybrid protocols, even without full error correction, and consider their role as a machine learning platform.