Is anyone working on wave-superposition-based pattern recognition instead of neural-network weights? [R]

A developer is exploring an alternative approach to low-level AI perception, focusing on wave-superposition-based pattern recognition rather than traditional neural network weights. This method aims to shift much of the early perception workload into the physical wave domain, potentially using physical logic gates for thresholds, routing, and nonlinear operations. The developer is seeking opinions from others in the field regarding this concept.