Julien Talbot — WritingEssays on real work, AI, agent delegation, supervision, and the conditions for trust in work systems.https://julientalbot.com/en-usThe Enterprise Agent Problem Is Beliefhttps://julientalbot.com/en/writing/enterprise-agent-problem-is-belief/https://julientalbot.com/en/writing/enterprise-agent-problem-is-belief/An agent is not enterprise-ready because it can act. It is enterprise-ready when the belief it creates about its action is calibrated to evidence.Thu, 04 Jun 2026 00:00:00 GMTAn AI talk should talk about real workhttps://julientalbot.com/en/writing/conference-ia-travail-reel/https://julientalbot.com/en/writing/conference-ia-travail-reel/Why a useful AI talk should start from what AI changes in work, decisions, responsibility, evidence and human recovery, not from tools alone.Tue, 02 Jun 2026 00:00:00 GMTRaw Traces Are Not Evalshttps://julientalbot.com/en/writing/raw-traces-are-not-evals/https://julientalbot.com/en/writing/raw-traces-are-not-evals/The missing layer between real agent failures and measurable model progress: reducing messy traces into replayable eval seeds without laundering the signal.Fri, 29 May 2026 00:00:00 GMTTool Use Is Not Task Completionhttps://julientalbot.com/en/writing/tool-use-is-not-task-completion/https://julientalbot.com/en/writing/tool-use-is-not-task-completion/Why agent reliability depends on preserving the boundary between intention, action, observable state, and truthful final claims.Tue, 19 May 2026 00:00:00 GMTThe Most Dangerous Agent Failure Is Not Hallucinationhttps://julientalbot.com/en/writing/claim-action-evidence/https://julientalbot.com/en/writing/claim-action-evidence/Why the critical risk is not only hallucination, but an agent claiming work is done without observable evidence.Thu, 14 May 2026 00:00:00 GMTAI Agents Are Not Just Tools. They Are Work Systemshttps://julientalbot.com/en/writing/agents-are-work-systems/https://julientalbot.com/en/writing/agents-are-work-systems/Why task completion is not enough: an agent redistributes verification, responsibility, coordination, and recovery work.Sun, 03 May 2026 00:00:00 GMTAI Is Not Integrated Into Work. It Reconfigures It.https://julientalbot.com/en/writing/ai-reconfigures-work/https://julientalbot.com/en/writing/ai-reconfigures-work/Why AI must be analyzed from real activity: constraints, trade-offs, invisible cooperation, room to act, and responsibility.Fri, 24 Apr 2026 00:00:00 GMTThe Benchmark Lie: Why Grok 4.20 Excels in Benchmarks but Fails in Productionhttps://julientalbot.com/en/writing/benchmark-lie/https://julientalbot.com/en/writing/benchmark-lie/A cognitive ergonomist dogfoods Grok 4.20 across 12+ production agents. What 40 years of human factors research says about what benchmarks miss in agent loops.Tue, 21 Apr 2026 00:00:00 GMT