Jensen Huang on OpenShell and AI safety
Jensen Huang discusses Nvidia’s OpenShell and Sentry systems for containing and monitoring AI agents, arguing that agent safety is an engineering problem based on minimal access rights and hardware isolation. He addresses liability, investment in OpenAI and Hugging Face, AI industry standards, regulatory timing, competition with China, model distillation, and Nvidia’s plans to return cash to shareholders.
Passages
0:45 “But we also have to make sure that the technology is developed and deployed safely so that the public has confidence. Applying this technology, the single most important thing we can possibly do is to build it in such a way that is so safe that everybody could use it in every single industry, so we could transform every industry.” Share
2:02 “Well, I think the answer is we hope it’s an engineering problem. I believe is an engineering problem.” Share
13:31 “Well, first of all, I think that we need to have an industrial set of standards and policies. And those industrial standards, like many other industries, set expectations of the level of quality.” Share
Search what Huang said from Claude, ChatGPT or your own code. See plans