Forget electrons, this breakthrough uses light-matter particles to power AI
AI Summary: Researchers at the University of Pennsylvania have developed a new approach to computing that utilizes light instead of electrons, addressing the limitations of electron-based hardware in modern AI applications. Led by physicist Bo Zhen, the team created a quasiparticle called an exciton-polariton, which allows for efficient signal switching necessary for computing tasks while consuming minimal energy—approximately 4 quadrillionths of a joule. This advancement could enable photonic chips to process information directly from cameras without the need for energy-intensive conversions between light and electricity, potentially reducing the power demands of large AI systems and supporting quantum computing functions in the future. The findings were published in *Physical Review Letters*.