Proton AI Paper Trail: What ChatGPT and Claude Really Know About You (and How to Find Out)

You’ve been using ChatGPT or Claude for months, maybe years. You’ve entrusted them with your project ideas, professional doubts, code snippets, and personal documents. But do you know exactly what these conversations reveal about you? The Swiss company Proton, known for its VPN and encrypted email, decided to make this question tangible with a free … Read more

Pentest: a methodical approach (step by step)

// pentest — a methodical approach (step by step) Penetration testing is not just an attack: it is a structured, planned process, and above all one framed by written authorization. This tutorial details the four phases of an engagement, with the tools and the reasoning behind each. A non-negotiable reminder: only act on systems you … Read more

Tech news roundup — week 33

// weekly tech watch Keeping up with technology is no longer done by hand. The RSS feeds from korben.info, Hacker News and The Hacker News are aggregated automatically in this site’s sidebar, offering real-time reading of the latest tech news. Why aggregate? Time saving — a single page instead of ten tabs. Cross-cutting view — … Read more

Open source: projects to watch in 2026

// ecosystem Open source is no longer a sub-sector: it is the foundation of the tech world. What we install today rests 90% on projects maintained by volunteer communities. Three families to watch Dev — lightweight runtimes, compilers, frameworks. Security — scanners, modular pentest tools. Infra & data — orchestration, databases, self-hosting. The diagram shows … Read more

Understanding LLMs: architecture and how they work

// language models A LLM (Large Language Model) is a neural network trained on enormous textual corpora. It does not “understand” in the human sense: it predicts the next token with a learned probability. The pipeline Prompt — input text. Tokenization — split into units. Layers — attention-based transformers. Logits — scores for the next … Read more

AI agents: toward autonomous assistants

// agentic An AI agent goes further than a chat: it acts. It combines an LLM, a reasoning loop and executable tools (shell, API, browser) to accomplish multi-step tasks without supervision at every step. Components Planner — breaks the objective into sub-tasks. Tools — functions that act on the system. Memory — persistent context between … Read more