NASA’s new dark energy space telescope can also detect killer asteroids
technologyreview.com·2h ago
Cloud & Infrastructure
The New Stack·Jun 29·5 min read
Palantir and Nvidia want to change who owns government AI
TL;DR
Government agencies face challenges using external AI models due to data security concerns. Palantir and Nvidia developed a framework for deploying and owning AI models securely within government networks.
✦ Why It Matters
Engineers can now explore secure AI deployment methods tailored for government and critical infrastructure applications.
Key Takeaways
- Assess the cost of owning AI infrastructure versus using hosted services.
- Explore the capabilities of Nemotron models for your specific needs.
- Implement a telemetry loop for continuous model improvement.
- Ensure your team is prepared for the operational demands of self-hosting AI.
How It Works
Palantir's intelligent engine allows organizations to deploy AI models within their own secure networks. It integrates Nvidia's open-weight Nemotron models, which utilize a mixture-of-experts architecture to optimize performance and reduce costs.
The system captures usage data to continuously improve model performance, ensuring that AI remains aligned with operational goals.