Sacha Storz – Blog in English https://storz.net/blog/en/ Thoughts on AI, work, and judgment by Sacha Storz. en-US Sat, 10 Oct 2026 00:00:00 GMT storz.net static blog builder Decision Models in My Personal News Filter: What Comparing Them with Traditional Classification Reveals https://storz.net/blog/en/decision-models-personal-news-filter/ https://storz.net/blog/en/decision-models-personal-news-filter/ Sat, 10 Oct 2026 00:00:00 GMT Three specialized Decision Models compete with a news classifier trained on personal choices. The experiment shows why semantic capability alone is not enough. Local AI Evals The Job Survives. But How Will Juniors Learn the Craft? https://storz.net/blog/en/job-survives-how-juniors-learn-the-craft/ https://storz.net/blog/en/job-survives-how-juniors-learn-the-craft/ Sun, 04 Oct 2026 00:00:00 GMT AI can take over tasks through which people build professional experience. How will they then learn to judge its results? Future Skills Who Reviews Every Output of Agentic Work? Does Anyone Have To? https://storz.net/blog/en/who-reviews-agentic-work-results/ https://storz.net/blog/en/who-reviews-agentic-work-results/ Wed, 23 Sep 2026 00:00:00 GMT Why human-in-the-loop cannot mean reading every AI output—and what risk-based controls for code and documents could look like. Agentic Work Evals Critical Reading: When the Backtest Gives the Agent Its Next Idea https://storz.net/blog/en/nvidia-finance-agent-backtest-feedback/ https://storz.net/blog/en/nvidia-finance-agent-backtest-feedback/ Fri, 18 Sep 2026 00:00:00 GMT NVIDIA's finance agent uses backtests to develop new signals. When is such an evaluation still an independent test? Critical Reading Evals My AI Coding Agent Kept Shipping. I Stopped Learning. https://storz.net/blog/en/ai-coding-agent-stopped-learning/ https://storz.net/blog/en/ai-coding-agent-stopped-learning/ Sat, 12 Sep 2026 00:00:00 GMT How I changed a productive agent workflow when project progress and my own understanding quietly drifted apart. Agentic Work Future Skills The Moxie case: If you design for attachment, you must also design for the end https://storz.net/blog/en/moxie-attachment-product-end-of-life/ https://storz.net/blog/en/moxie-attachment-product-end-of-life/ Mon, 07 Sep 2026 00:00:00 GMT AI companions can create stronger personal bonds than ordinary apps. That also increases the responsibility for their lifecycle management. Critical Reading Future Skills Andrew McAfee on the AI era: What makes organizations truly capable of learning? https://storz.net/blog/en/what-makes-organizations-capable-of-learning-in-the-ai-era/ https://storz.net/blog/en/what-makes-organizations-capable-of-learning-in-the-ai-era/ Mon, 31 Aug 2026 00:00:00 GMT The winners in the AI era may not be the flattest or fastest organizations, but those that turn new evidence into reliably better decisions. Critical Reading Agentic Work TrueForge boasts 30%-75% cheaper task completion - but what did they really test? https://storz.net/blog/en/trueforge-30-75-percent-cheaper-what-did-they-test/ https://storz.net/blog/en/trueforge-30-75-percent-cheaper-what-did-they-test/ Fri, 28 Aug 2026 00:00:00 GMT A number like “75% cheaper” does not tell us where the efficiency comes from. The TrueForge benchmark shows why we should distinguish the harness from the system as a whole. Critical Reading Agentic Work Evals