Featured Two security deals for AI agents this week, and only one is a deal Five hundred million dollars in cash, plus payments tied to milestones: that is what Palo Alto Networks would pay for… Nov 9 285K views 4 min read
Featured Mistral raises €3 billion Series D led by Samsung Electronics Mistral has raised €3 billion in a Series D round at a post-money valuation of more than €21 billion, led… Nov 15 253K views 4 min read
Featured Has the Cohere and Aleph Alpha merger closed? Not yet. Cohere and Aleph Alpha, the German AI company, signed a definitive merger agreement on 16 September 2026, and… Nov 11 146K views 3 min read
When a Hallucination Is a Design Constraint, Not a Bug A memory leak is a property of how programs are. A race condition is a property of how concurrent programs… Jul 25, 2026 Yves Mulkers 39K views 12 min read
Reducing hallucinations: what the response layer actually does The first time I watched a team try to “fix hallucination” with a single technique, they spent three months tuning… Jul 24, 2026 Yves Mulkers 40K views 13 min read
A Taxonomy of Hallucinations: Intrinsic, Extrinsic, and the Systemic Kind You cannot mitigate a failure mode you cannot name. That sentence sounds obvious until you sit through a post-incident review… Jul 23, 2026 Yves Mulkers 92K views 11 min read
Continuous Evaluation: When the Benchmark Is Your Production Traffic The held-out test set was a useful fiction. It got us through the first wave of putting language models into… Jul 22, 2026 Yves Mulkers 96K views 12 min read
Evaluating RAG Systems: Faithfulness, Answer Relevance, Context Precision The first useful thing to say about evaluating a retrieval-augmented system is that you are not evaluating a model. You… Jul 21, 2026 Yves Mulkers 199K views 11 min read
LLM Evaluation: From Leaderboard Scores to Harness Proof The leaderboard tells you the engine works in a vacuum. The harness evaluation tells you the system holds in your… Jul 20, 2026 Yves Mulkers 110K views 11 min read
Context Window Management: Truncation, Eviction, and the Failure Mode When the Model Runs Out of Room The first time I caught this failure in the wild, it was in a long customer-support conversation. The agent answered… Jul 19, 2026 Yves Mulkers 84K views 12 min read
Context as a Guardrail: System Prompts, Few-shot, Retrieved Context, Tool-Call Context There is a guardrail nobody on the call ever names. The team will argue for an hour about output schemas,… Jul 18, 2026 Yves Mulkers 122K views 12 min read
Output Handling and Structured Response Gates: Keeping a Wrong Answer From Becoming a Wrong Action The first production incident I watched cleanly involved an agent that did not hallucinate. The model returned a perfectly fluent… Jul 17, 2026 Yves Mulkers 81K views 12 min read
Prompt Injection and the OWASP Top 10 for LLM Applications Most teams I sit with have the same picture of prompt injection, and most are looking at the wrong picture.… Jul 16, 2026 Yves Mulkers 30K views 13 min read