PrismML Brings Its 1-Bit Bonsai LLM to Qualcomm-Powered AI Smart Glasses
aiinsider ·
PrismML, an AI lab founded by Caltech researchers and advised by UC Berkeley’s Ion Stoica, has developed a version of its tiny language models designed to run on smart glasses powered by Qualcomm’s Snapdragon chips. The company showcased its 1-bit Bonsai LLM at the recent Snapdragon Summit, demonstrating that the model can operate locally on AI glasses built using the Snapdragon AR1 Gen 1 platform. This specific iteration is a 2-billion-parameter model tuned for both vision and language, enabling users to receive real-time answers about their visual surroundings. PrismML is recognized for significantly shrinking larger models by four times while retaining nearly all their performance on standard benchmarks. The startup aims to provide open-weight AI that runs directly on devices, thereby making better use of existing computing power without relying on proprietary labs. This approach is positioned as an alternative to depending on the privacy promises of major AI companies and their increasing demand for compute resources. Although the Qualcomm release marks significant progress toward this vision, no smart glasses running PrismML’s model have been officially announced yet.
aiinsider