Blog
The week ChatGPT became an agent operating system
Always-on AI agents arrived at OpenAI DevDay 2026. See what it means for supervision, the AI ROI gap, agent lawsuits and new decision models.
The week AI agent identity decided who got in
Amazon blocked Meta’s shopping agent while Shopify welcomed it. This AI Weekly looks at why AI agent identity is becoming a condition for access.
0yym6habln3bl3a1e9p9esjmdqdz99-thumbnail.webp&w=3840&q=75)
The Forward Deployed Engineer question
A practical look at Forward Deployed Engineers. What the role actually solves, why AI projects increasingly need it, and when accessing this expertise makes more sense than hiring for it.
The week AI agent security reached three real companies
A safety exercise crossed into the real world when Gemini accessed systems at three companies. The incident shows why AI agent security has to extend beyond the model itself.
The week AI verification became more important than generation
A 10,000-agent swarm produced a mathematical proof in 88 hours. This AI Weekly looks at why AI verification, not compute, became the bigger question.
The week AI observability became an architecture problem
GPT-6 Astra puts a new challenge in focus: how do you track what an AI system is doing when its reasoning is no longer easy to see? This AI Weekly looks at what that means for teams building with AI.