Journal · one note per week
Raw notes on AI in production, real agentic systems, and what actually works for the startups we work with.
↳ 5319 notes
Terminal-Bench-Science 0.1 establishes a new benchmark for evaluating AI agent capabilities in scientific research workflows. Developed by Stanford, this benchmark involves scientists in setting evaluation criteria, encompassing 70 tasks from various scientific fields.
ReadThe German Sovereign Tech Agency has invested over €500,000 in Flatpak, aiming to enhance the security and user experience of Linux desktop software.
ReadNitter, an open-source alternative to Twitter, was forced to shut down following a cease-and-desist letter. Discover why and what this means for the tech ecosystem.
ReadLearn how to harness Jupyter notebooks on Google Colab to develop AI engineer skills without relying on complex frameworks. Explore concrete use cases of RAG, agents, and evaluations.
ReadBiotechnology is revolutionizing the food industry by converting plastic and biomass waste into food additives using engineered yeast.
ReadStripe has recently decided to abandon its $50 billion bid to acquire PayPal. Let's delve into the reasons behind this decision and its implications for the digital payments sector.
ReadThe new Model Hardware Standard (MHS) aims to transform how AI agents interact with physical devices. Get ready for quick and efficient integration of lab and manufacturing instruments.
ReadVoronoi Go blends the beauty of Voronoi diagrams with the complexity of the game of Go. Discover how this innovation is transforming AI learning and inspiring new applications.
ReadM5Stack launches PaperMono, a compact ESP32-S3-based development terminal with an E-Ink display, perfect for energy-efficient IoT projects.
Read30 minutes, free. You leave with three concrete actions, even if we don’t work together.