#shorts #Nvidia #TensorRT #generative AI #LLMs #GPUs #inference
This just in: Nvidia is making moves to extend its lead in generative AI. Amyovich reports the computing powerhouse released software to accelerate large language models like GPT-4. Its new TensorRT toolkit aims to speed up AI inferencing on Nvidia’s pricey GPUs. This could encourage adoption despite the high hardware cost. Jostovich says Nvidia wants to own both the lucrative training and inferencing sides of generative AI. But new chips from Microsoft and AMD could reduce reliance on Nvidia. Still, its aggressive push could cement its central role as creative AI takes off.