Q&A: Nvidia genAI chief explains why open models matter in AI
When Nvidia CEO Jensen Huang speaks, the tech industry listens. He used his first-ever post on X last week to argue that open AI models “strengthen safety and cybersecurity, accelerate innovation and diffusion, and enable sovereignty.” Nvidia is a chip company (with a focus on GPUs). But it also has its own AI models that include Nemotron, an open-weight model that’s free to download and modify.…
Nvidia CEO Jensen Huang recently expressed his views on open AI models in a post on X, emphasizing that they "strengthen safety and cybersecurity, accelerate innovation and diffusion, and enable sovereignty." Nvidia, a leading chip company with a focus on GPUs, also develops its own AI models, including the open-weight model Nemotron. Kari Briski, Nvidia's vice president of generative AI software, discussed the importance of open models and why they matter for enterprises and sovereign applications.
Open models offer enterprises and countries the ability to own, inspect, and adapt models using their own data. This leads to faster adoption and progress in closing the gap between different regions and enterprises. Nemotron, published by Nvidia, also shares its data, which has opened up new opportunities for engagement with enterprises interested in curating, creating, capturing, and controlling their own data.
Briski noted that the target audience for open models encompasses both CIOs and sovereign uses, with a focus on use cases, data sets, and outcomes. Enterprises may have thousands of tools, while local regions may have thousands of regional tools, but the fundamental goal remains the same. This allows Nvidia to tailor post-training on local models to match local ecosystems.
Nvidia releases its models in the open to foster a thriving ecosystem, enabling startups, builders, and lower entry barriers. By sharing their models, Nvidia gains valuable insights and benefits from the contributions of the broader community. Open models and open data act as a bootstrap for regions lacking the necessary compute power to achieve a base-level model. This is especially crucial for regions like South Asia, where voice and word of mouth are more prevalent than chatbots.
Briski emphasized that the approach to AI deployment varies across regions and depends on factors such as language evolution and the development of a data flywheel. The goal is to continuously update and adapt AI models in everyday applications, turning AI into a new computing platform. Nvidia offers different sizes of Nemotron models, ranging from Nano to Ultra, catering to various deployment scenarios and developer needs.
The company also collaborates with cloud providers and inference partners to ensure efficient and state-of-the-art models are accessible to users.
Written by urgent.news from Computerworld's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
