Posts
All the articles I've posted.
-
GPT-5.6 Shows Frontier Access Becoming Conditional
OpenAI's reported GPT-5.6 limited rollout is another sign that paid access to closed frontier AI may no longer mean equal access to the frontier.
-
GLM-5.2 Is the Counterargument to Gated Frontier Access
Z.ai's GLM-5.2 release matters because open weights, a 1M-token context, and an MIT license make frontier dependency less fragile for builders.
-
OpenAI's Jalapeno Chip Is About Owning the Whole Stack
OpenAI's first custom Broadcom inference chip shows frontier AI labs becoming full-stack infrastructure companies, not just model providers.
-
GitHub Put AI Credit Burn Into The Copilot Metrics API
GitHub's per-user AI credit metric shows agent cost control is becoming an enterprise observability problem, not just billing noise.
-
DeepMind's AI Control Roadmap Treats Agents Like Insider Risk
Google DeepMind's AI Control Roadmap is a signal that agent safety is moving from alignment slogans into security architecture.
-
A2A And ARD Are Turning Agents Into A Network
Google's A2A anniversary and Hugging Face's ARD launch show agent infrastructure moving toward discovery, delegation, and interoperable peers.
-
MosaicLeaks Shows Research Agents Can Leak Without Saying Secrets
Hugging Face and ServiceNow's MosaicLeaks benchmark shows deep research agents can leak private facts through ordinary web searches.
-
OpenAI's Rare-Disease Study Shows The Right Shape For Medical AI
OpenAI's June 18 health work shows why medical AI is strongest as expert-led reanalysis, not unsupervised diagnosis theater.
-
GitHub Agent Finder Is About Tool Discovery, Not More Chat
GitHub's Agent Finder and Copilot governance updates show the next agent problem: finding the right tool without losing enterprise control.
-
Microsoft Is Making Agent Cost A Product Feature
Microsoft's June 16 AI post frames model diversity, Agent 365, and Copilot Cowork usage pricing as the enterprise cost layer agents need.