Richard Wang has successfully defended his thesis titled Physics-inspired computing and its applications. He enjoyed a well deserved post-viva punting trip on the River Cam with his fellow group mates.

In our recent Nanophotonics paper (Peng et al., 2024), we use a hybrid organic–inorganic perovskite, a semiconductor studied for solar cells, to build room-temperature lattices that mimic interacting magnetic spins. Our building block is a polariton condensate: light trapped between two mirrors couples strongly to the crystal, forming part-light, part-matter particles that, above a threshold laser power, act as one wave. Using patterned laser light, we place condensates in pairs and squares of four, whose neighbours synchronise in step or exactly out of step, corresponding to aligned or opposed spins. Widening a pair's spacing from 1.78 to 3.55 micrometres switches it from opposed to aligned and back. Our analysis shows these lattices realise a soft-spin XY model, whose spins can point anywhere in a plane and vary in strength. Such lattices could simulate complex magnetic materials and solve optimisation problems; building larger ones remains future work.

Conventional chips are hitting hard limits on speed, heat and energy use. In their our review, published in Advanced Quantum Technologies, we explore a promising alternative: computing with light. Rather than performing operations numbers step by step, networks of lasers or light–matter particles called polaritons let physics do the work, naturally settling into the state that loses the least light, which is designed to encode the answer to a hard problem like planning a delivery route or balancing a portfolio. The review covers the field’s history, the mathematics these devices share, and uses ranging from finance to AI, and argues that light is more likely to become a specialised accelerator than a replacement for electronics.