Nvidia launches G-Assist: Local AI Assistant provides full support for players

Nvidia recently launched Project G-Assist, an experimental AI assistant, which runs locally on a GeForce RTX card. G-Assist relies entirely on the user's GeForce RTX GPU, and uses the 800 million parameter model based on Llama, and is activated by pressing the Alt+G key. The assistant has functions such as system diagnosis, game optimization, GPU overclocking and performance monitoring, and can control compatible peripherals. Nvidia also designed community extensions for G-Assist and provided a GitHub code base to enable developers to create plugins. G-Assist is not Nvidia's first launch of a local AI assistant, and it has previously updated the experimental chatbot ChatRTX and launched the Nemotron-44B Instruct language model. This development is in line with Nvidia's vision to deeply integrate AI systems with gaming and other applications.