another deal with Uncle Sam. This time, it will get its latest models into the hands of US scientists working on nuclear safety and more.
Microsoft-backed AI maker OpenAI today announced a US National Lab partnership. The partnership will begin with the deployment of its reasoning capable o1 LLM on Los Alamos National Laboratory (LANL) Venado supercomputer . OpenAI hopes that the Nvidia-powered beast will supercharge their model and enable various scientific breakthroughs.
OpenAI says that LANL plans to use o1 for a variety of purposes, including identifying new approaches to treating disease, protecting the US power grid and enhancing America’s general and cyber security. It will also contribute to high energy physics research.
This announcement also plays to President Trump’s previous executive order regarding US energy leadership. Citing o1 at LLNL, the announcement cites “unlocking the full potential of natural resources and revolutionizing the nation’s energy infrastructure” and “accelerating the basic science that underpins US global technological leadership.”.
o1, just like the headline-grabbing DeepSeek model which came out of China, is a chain of thought model that’s meant to hallucinate fewer times and spit better answers
Venado o1 will be available to researchers at Lawrence Livermore National Laboratory, Sandia National Laboratory, and other labs. OpenAI stated that the three labs would collaborate on a program aimed at reducing the risk of nuclear warfare and securing nuclear weapons and materials around world. ChatGPT said
“This use case is highly consequential, and we believe it is critical for OpenAI to support it as part of our commitment to national security,” and added: “Our partnership will support this work, with careful and selective review of use cases and consultations on AI safety from OpenAI researchers with security clearances.”
OpenAI has worked with LANL in the past. The pair collaborated in order to investigate the use AI in bioscientific researchers. OpenAI said that the new partnership builds on the previous program, which involved the use of GPT-4o to assess the risk that artificial intelligence could be used to develop bioweapons.
- Is AWS able to fix AI hallucination problems? We speak to Byron Cook, head of Automated Reasoning
- OpenAI’s Operator agent wants your online chores – but don’t expect it to nail them all.
- OpenAI is looking to spend $500B on AI infrastructure, with the help of friends
- DeepSeek hasn’t finished yet with OpenAI – image-maker Janus Pro wants to take on DALL-E 3!
OpenAI didn’ It’s funny that Microsoft, who helped fund OpenAI, promised that they would distribute the model free to all Copilot customers.
This is the second announcement OpenAI has made this week regarding working with the US Government. The biz announced on Tuesday that ChatGPT Gov was released, a variant designed for the public sector. OpenAI, which has had its products tested in several state governments and federal organizations prior to the release of ChatGPT, hopes that the new product will expedite the process to be authorized to manage “non-public sensitive data,” an issue it is currently working on. (r)
