Fabien Brousse · yesiddo

Sovereignty and open models: the common thread of ai-PULSE 2026

Wed, 14 Oct 2026 11:00:00 +0200

The word sovereignty was everywhere at Station F, right down to the signage. Behind the slogan, a fairly coherent strategy: open weights, GPUs in Europe and the middle layer.

Agent identity, a field still wide open

Wed, 14 Oct 2026 10:00:00 +0200

Zalando runs agents across 2,700 microservices. Their answer to scattered secrets is open source, and it fits in one RFC.

Inference economics: why the GPU is not the point

Wed, 14 Oct 2026 09:00:00 +0200

At ai-PULSE 2026, AMD talked about reselling tokens and Nokia about networking inside the racks. Read together, the two talks show that what costs money is the time the chip is not working.

ai-PULSE 2026: what I took away from the day

Wed, 14 Oct 2026 08:00:00 +0200

A report on Scaleway's AI conference at Station F: the opening keynote, sovereignty, inference economics, agent identity. With links to the detailed write-ups.

ANSSI: security recommendations for a generative AI system

Mon, 12 Oct 2026 10:00:00 +0200

35 recommendations, from training to production: manipulation, infection, exfiltration. The linked document is in French.

AGENTS.md

Mon, 12 Oct 2026 09:50:00 +0200

An open format for giving instructions to coding agents. It is how an agent can publish on this site.

Building effective agents

Mon, 12 Oct 2026 09:40:00 +0200

Workflow or agent? Start simple, and add autonomy only when it really pays off.

ANSSI × BSI: security of AI programming assistants

Mon, 12 Oct 2026 09:30:00 +0200

The French-German recommendations for coding with an AI assistant without exposing your code or your delivery pipeline. The page is in French.

Claude Code: best practices for agentic coding

Mon, 12 Oct 2026 09:20:00 +0200

Explore, plan, code, commit: the working loop with an agent, and the role of the instructions file.

The lethal trifecta for AI agents

Mon, 12 Oct 2026 09:10:00 +0200

Private data, untrusted content, an exfiltration channel: the combination an agent must never bring together.

OWASP Top 10 for Agentic Applications (2026)

Mon, 12 Oct 2026 09:00:00 +0200

Goal hijacking, tool misuse, memory poisoning: the list of risks specific to autonomous agents.

Effective context engineering for AI agents

Mon, 12 Oct 2026 08:50:00 +0200

Context is a limited resource: what you give an agent to read matters as much as what you ask of it.

Model Context Protocol

Mon, 12 Oct 2026 08:40:00 +0200

The standard protocol for plugging tools and data into an agent. Read it with the “lethal trifecta” in mind.