When the precedent hasn’t been set yet, we get to write it
The accountability gap doesn't just create compliance risk. It creates operational security risk. When model developers point to deployers and deployers point to model developers, the space between them becomes the attack surface.
A new research paper from Stanford, Harvard, UC Berkeley, and Caltech — "Adaptation of Agentic AI" — provides the clearest framework I've seen for diagnosing what goes wrong when agentic AI systems move from controlled demonstrations to real-world deployment.
Before MCP, every AI application needed custom connectors for each data source. Without foundation governance, that success creates three risks: proprietary lock-in, protocol fragmentation, or de facto control by a single company. AAIF prevents all three
Advancing AI Negotiations: New Theory and Evidence from a Large-Scale Autonomous Negotiations Competition Authors: Michelle Vaccaro, Michael Caoson, Harang Ju, Sinan Aral, and Jared R. Curhan
You can't eliminate non-determinism in LLMs, and you shouldn't try. The goal is management, not elimination.
Most teams building agentic AI are trying to make LLMs do something they're fundamentally bad at: making consistent, defensible business decisions. Decision platforms were built for this exact problem - and most teams don't know they exist.
At its core, AI red teaming is a multidisciplinary, creative, and interactive process of investigation
Headlines proclaim that AI will change everything overnight—jobs, society, the whole works. The technology is powerful, no question. But some experts want us to slow down and look past the hype.
