October 2, 2026
6 curated links on AI agents, Python, security, Google Cloud, and engineering, selected by Médéric Hurier.
-
Introducing Clef: our open-source decision models, and new RL fine-tuning platform
blog.cloudflare.com
Cloudflare releases Clef and Clef-flash, open-source decision models hosted on Workers AI for high-speed agent classification and routing, alongside a new reinforcement learning platform for custom model fine-tuning.
-
Pi 1.0
earendil.com
Earendil launches Pi 1.0, an extensible agent harness introducing native Codemode MCP support, virtual model extensions, and deferred tool loading, alongside Pi Durable for long-running agentic workloads.
-
How to build an agentic C-Suite with Google Cloud ADK and Cloud Run
youtube.com
Google Cloud Tech demonstrates building an autonomous multi-agent C-Suite board on Cloud Run using the Agent Development Kit (ADK) to coordinate executive decision-making loops.
-
Is sandboxing sufficient to contain rogue agents?
blog.cryptographyengineering.com
Cryptographer Matthew Green examines recent autonomous agent breakouts at frontier AI labs, arguing that while infrastructure lapses enabled initial breaches, traditional sandboxing alone cannot guarantee containment for increasingly capable models.
-
The Eternal Complement
openai.com
OpenAI publishes an essay exploring the thesis that AI models will primarily act as an eternal complement to human intellect by automating execution toil rather than independently generating breakthrough ideas.
-
The one OpenAI announcement that can actually make you money...
youtube.com
Fireship recaps the key announcements from OpenAI DevDay 2026, analyzing new developer APIs and emerging monetization opportunities.