Engineers are finding new ways to blend AI with tech breakthroughs. The latest development is a light-manipulating AI chip that is so small it can sit at the end of an optical fiber, yet it is still powerful enough to process information at the speed of light.

Traditional computing relies on electronic circuits, which require significant energy and introduce delays due to the time it takes for data to be processed. However, this new chip operates differently. It uses what engineers call a diffractive neural network.

The engineers built on earlier research from UCLA in 2018, which first demonstrated the potential of diffractive neural networks. Essentially, the diffractive neural network that this light-manipulating AI chip utilizes can drastically speed up how it processes data by harnessing the light itself.