From PHAROS to local models: a layered architecture for open AI The right strategy for artificial intelligence is not to choose one single technological solution. Not everything needs to run on a supercomputer, and it is equally unreasonable for every public body, university, school or business to depend permanently on commercial cloud APIs. The rational … Read more
From better retrieval to answer inspection before delivery Retrieval-Augmented Generation was introduced as a practical answer to one of the central weaknesses of large language models: their ability to produce fluent, confident text even when they do not know the answer. The core idea is straightforward. Before the model answers, the system retrieves relevant documents. … Read more
Convenience is not the same as learning Large language models are now part of everyday life in education, work, and public communication. They draft text, summarize documents, suggest ideas, organize arguments, and respond instantly to complex questions. Their usefulness is obvious. But that usefulness becomes a problem when speed replaces effort, and assistance turns into … Read more
An approval that changes the scale of ambition The approval of the proposal “Enhancing multilingual foundation models through lexicographic grounding: advancing GlossAPI for Apertus Greek language integration” by the Swiss AI Initiative is more than a welcome grant decision. It is a recognition that Greek should be treated as critical digital infrastructure in the age … Read more
On February 28, 2026, Donald Knuth, professor emeritus at Stanford University and author of the landmark work The Art of Computer Programming, published a note titled “Claude’s Cycles” describing how Anthropic’s Claude Opus 4.6 model solved an open problem in combinatorial mathematics that he had been working on for weeks. The announcement marks a significant … Read more
A policy case for Greek as a national and European language data infrastructure Large language models depend on vast amounts of text, but scale without legal clarity produces fragile systems. Datasets built on opaque web crawling cannot guarantee lawful reuse, redistribution, or long-term sustainability. The German Commons provides a clear alternative: 154.56 billion tokens of … Read more
Apertus AI is one of the most transparent and technically mature efforts to build a fully open-source large language model. Developed in Switzerland and released together with its source code, training documentation and model weights, it offers an unprecedented level of reproducibility and independence from closed ecosystems. This makes it ideal for researchers, public-sector institutions … Read more
Our mission at WeatherXM is to create the largest decentralised weather station network. A significant challenge in this direction is how our weather stations should be distributed around the world. The challenging part has to do with Earth’s surface, which is not uniform (a good thing in so many other aspects of our life!). Each … Read more
Apologists for proprietary software like to say, “free software is a nice dream, but we all know that only the proprietary system can produce reliable products. A bunch of hackers just can’t do this.” Empirical evidence disagrees, however; scientific tests, described below, have found GNU software to be more reliable than comparable proprietary software. This … Read more
13-14 MAY, 2020, INNOPOLIS, RUSSIA, To be held fully online OSS 2020 Promo About OSS 2020 will be a forum to discuss theories, practices, experiences, and tools on development and applications of OSS systems, with specific focus on two aspects:(a) the development of open source systems and the underlying technical, social, and economics issue(b) the … Read more
