-
-
Prompt injection at work: hidden instructions in emails and documents
Many offices now let Copilot, ChatGPT, Gemini or Claude summarise the inbox, pull the key figures out of an attached invoice or draft a reply. To do that, the language model reads everything in the email and the attachment, including parts you never see. If that content contains an instruction, the model may follow it…
-
-
-
Model Context Protocol (MCP): connecting AI to tools and data
Every AI application that needs to reach a database, a helpdesk or a file store has traditionally needed its own integration for each of them. With several applications and several systems, the number of connectors multiplies quickly. The Model Context Protocol, MCP, replaces that with a single open protocol: one server has to be built…
-
Agentic AI: systems that plan and act on their own
The word agent is attached to almost every AI product today, which makes it easy to lose sight of what it means. Agentic AI describes systems in which a language model does not just answer but directs its own process: it breaks a goal into steps, chooses and uses tools, looks at the results and…
-
Red teaming AI systems: finding the weak spots on purpose
Imagine a customer service chatbot that has passed weeks of testing: it answers questions about delivery times, returns and invoices correctly. Then someone asks it to ignore its instructions and reveal them, and it does. The chatbot had been tested for everything it was supposed to do. Nobody had tried to make it misbehave, and…
-
Fine-tuning: retraining an AI model on your own examples
The first question about fine-tuning is not how, but whether. Fine-tuning means taking a model that has already been trained and training it further on your own examples, so that it performs a specific task, format or tone more reliably. Google’s glossary calls it a second, task-specific training pass on a pre-trained model. Three ways…
-
Embeddings: how AI turns meaning into numbers
An online shop sells a “rain shell”. A customer searches for “waterproof jacket” and finds nothing, because the words do not match. An embedding fixes exactly this kind of miss: it turns a piece of text into a list of numbers, a vector, and texts that mean similar things receive similar vectors. The search then…
-
Claude Code worktrees: parallel sessions without collisions
Two Claude Code sessions in the same directory are a recipe for trouble: one builds a feature, the other fixes a bug, and both write to the same files. Git worktrees have solved this for humans for years, and Claude Code has the route built in. This article explains how to start a session in…
