This post is also available as a video on our YouTube channel AI Crime Stories from our SRH Distance Learning University: Can we make large language models more humble? 🤖 The AI crime thriller about megalomania. Contents Overconfidence in humans Overconfidence in large language models Why do LLMs overestimate themselves? The Unholy Alliance Can We…
The French economic fabric—and its SMEs in particular—is structurally vulnerable to reputational attacks. And almost no one is taking this seriously enough to prepare for it. Subscribe to Must-Read News On Wednesday, April 8, at the École Militaire, the Directorate General of Armaments hosted a presentation of the findings from theEconomic Intelligence in Defense—the RIDES….
In this blog post, I want to show you how I create my prompts largely automatically. I’m focusing here on prompts for complex tasks and workflows. For simple tasks, you no longer need sophisticated prompt engineering these days. Large Language Models (LLMs) are now so good that they can produce good outputs even with poor…
It’s a parking rule that causes a lot of trouble for candidates taking the written driver’s license exam. Subscribe to Must-See News Filmed scenarios, eco-driving, new technologies: since 2016, the written driver’s license exam has undergone a major overhaul. Hundreds of new questions have been created to populate the exam’s 40-question pool. However, when the…
This post follows up on the penultimate post on this blog Can we make large language models more modest? . Both posts are also available as videos on the YouTube channel AI Crime Stories of our SRH Distance Learning University. But why should AI agents be at risk? Five key reasons can lead to the…
Be careful with your credit card. At gas stations, a discreet trap can clone your card and cause unexpected charges to your account. Subscribe to Must-Read News This is the story of an invisible heist that costs Americans nearly a billion dollars every year. The thieves don’t target banks or armored trucks. They operate in…
On this question and the new promptingtechnique 𝗩𝗲𝗿𝗯𝗮𝗹𝗶𝘀𝗶𝗲𝗿𝘁𝗲𝘀 𝗦𝗮𝗺𝗽𝗹𝗶𝗻𝗴 (𝗩𝗦), we have a fresh, detailed video on the YouTube channel of the SRH Distance Learning University – The Mobile University K. I. Crime Novels uploaded: 𝙆.𝙄. 𝙆𝙧𝙞𝙢𝙞𝙨: 𝙂𝙞𝙗𝙩 𝙚𝙨 𝙙𝙚𝙣 𝙢𝙖𝙜𝙞𝙨𝙘𝙝𝙚𝙣 𝙋𝙧𝙤𝙢𝙥𝙩? Verbalized sampling was introduced in the study How to Mitigate Mode Collapse and…
The CMO, often confined to operational duties, struggles to establish themselves as a strategic player. Their legitimacy depends above all on their approach: speaking the language of business, collaborating internally, and demonstrating strong leadership. Subscribe to Must-Read News More and more Chief Marketing Officers (CMOs) now sit on executive committees. Yet, in many tech companies,…
Since we enjoy conducting academic research, I’d like to introduce you to the best AI tool for the job: NotebookLM by Google. As of late March 2026, the base model behind NotebookLM is Google’s Frontier model Gemini 3.1 Pro. In this post, I’ll explain when and for what purposes we use NotebookLM versus when we…
After years of assistants and co-pilots that could only make suggestions, our deepest wishes seem to have come true: An AI assistant to whom we can describe our computer problem in natural language and who fixes the problem immediately—instead of sending us to countless forums to find solutions: Claude Code doesn’t make suggestions. Claude Code…