Agentic AI Security Is 10x Harder Than LLM Safety
Why the security surface of an agent system is categorically different from a model’s, and what that means for the open stack.
From Code to Agency: Open Source and Standards for the Next Generation of AI Models and Agents
Where the open stack has to go as models stop generating and start acting — and which standards have to exist first.
The Good, the Bad and the Ugly of Open Source AI
Open source is shaping the future of artificial intelligence — and not all of it is good. What openness actually buys us, where the term is being abused, and what the community should do about it.
No Agents Without Standards
Why an agent ecosystem without shared runtime standards cannot be audited, reproduced or trusted.
Open AI (Two Words): The Only Path Forward for AI
The case that open source and open science — two words, not one company — are the only durable path for the field.
Model Openness Framework: The Path to Openness, Transparency and Collaboration in Machine Learning
The case for the MOF: seventeen lifecycle components, three classes of openness, and a way to say what a model release actually includes.
Open Source AI and PyTorch
The foundation’s position on open source AI, and where PyTorch fits in it.
Model Openness Framework: The Path to Openness, Transparency and Collaboration in Machine Learning
Presenting the MOF with Anni Lai — seventeen lifecycle components, three classes of openness.