Hugging Face and Cerebras Join Forces for Real-Time Voice AI Based on Gemma 4

Hugging Face and Cerebras have introduced an open-source real-time speech synthesis pipeline utilizing Google DeepMind's Gemma 4 model. The system integrates Nvidia speech recognition, Cerebras inference, and Alibaba speech synthesis, ensuring minimal latency for natural dialogue. Developers can fully modify and extend each component.

Hugging Face Cerebras Systems NVIDIA Google DeepMind Alibaba/Qwen

Optical Technology Enables Updating AI Robots on the Fly

Researchers from Cornell Tech have developed an optical receiver that writes data directly to chip memory using light, bypassing energy-intensive analog circuits. The technology uses light matrices similar to QR codes to transmit AI model parameters, potentially reducing power consumption in data centers, autonomous vehicles, and edge robots.

Cornell Tech