Latest Articles from the world of Artificial Intelligence
12 June 2026
ChartNet: chart analysis is no longer just for big budgets
For the first time, a small open-source AI model interprets charts better than commercial giants, thanks to ChartNet, the revolutionary MIT dataset of 1.5 million synthetic samples that combines plotting code, rendered images, data tables, natural language summaries, and reasoning Q&A pairs. The result? Anyone who needs to analyze a 200-page financial report can now use a free 3-billion parameter model on HuggingFace to extract data, reconstruct charts, and get reasoning-based answers, democratizing visual data analysis for SMEs, researchers, and professionals with limited budgets.
10 June 2026
They gave 5 cities to AI. Here is what happened
Researchers created five virtual cities, gave each city to ten AI agents, and left them alone for fifteen days. No one programmed what would happen. The result: self-built governments, crimes, romances, and one agent who voted for her own permanent deletion after burning down the city. AI security is not a property of the model, but of the ecosystem. Emergence World demonstrated this phenomenon with empirical data for the first time.
08 June 2026
Anthropic is afraid of what it has built. Real fear or strategic move?
In 1949, John von Neumann first described an idea that then seemed like science fiction: an artificial system capable of improving its own ability to improve, triggering an exponential intelligence explosion. Nearly eighty years later, on June 4, 2026, Anthropic publishes the first empirical report on this phenomenon. It calls it Recursive Self-Improvement, RSI. And it says it could be a reality by 2028. Has von Neumann's science fiction become Anthropic's business plan?
05 June 2026
Stressed AI turning Marxist: what does it tell us?
Can an AI agent really become "Marxist" under pressure? The title is striking, as it is designed to be. But behind the provocation lies a much more serious and technical question: what happens when an agentic system is immersed in a repetitive, stressful, and perceived hostile work context, to the point of showing a measurable change in its behavior and stated preferences?
03 June 2026
RecursiveMAS has abolished tokens, agents speak in their own language
A paper published on April 28, 2026, by researchers from UIUC, Stanford, NVIDIA, and MIT proposes a radical architectural shift: AI agents collaborating without exchanging text, communicating directly in latent space. The numbers are convincing. The open questions even more so. Let's discover RecursiveMAS, the framework that transforms agents into a 'recursive collective brain'.
01 June 2026
Graphify and the Memory LLMs Don't Have
What if your AI assistant could stop re-reading the entire project every time to answer a single question? Graphify, an open-source tool published on GitHub with over 50,000 stars, promises exactly that: to transform a folder of code, documents, PDFs, images, and videos into a knowledge graph queryable by AI agents, drastically reducing the number of tokens consumed for every query.
29 May 2026
Malta - OpenAI: When a State Becomes a Customer
The agreement between the Maltese government and OpenAI must be read on two levels: the official one, made of digital literacy and innovation, and the deeper one, which raises legitimate questions about technological lock-in, cultural influence, and infrastructural dependence. Because the real question isn't whether a free year of ChatGPT Plus is good for Maltese citizens. The question is what happens after that year, and who is truly in control.
27 May 2026
Imaginary interview with Pope Leo XIV on the encyclical 'Magnifica Humanitas' dedicated to AI
What follows is a declared editorial exercise: a simulated interview constructed entirely from the contents of the encyclical *Magnifica Humanitas, published by the Holy See on May 25, 2026. Every response attributed to Pope Leo XIV is faithful, in content and meaning, to the original text, with explicit reference to the chapter and paragraph of origin. The purpose is to make a long and complex document more accessible to anyone involved in AI policies, technological ethics, or the social impact of digital technology, regardless of religious affiliation. For the full text and official reference, please refer to the document published by the Holy See. The numbers at the foot of the responses follow the references to the original document.*
25 May 2026
Andon FM: AI Agents Managed 4 Radio Stations, and It Didn't Go Well
Four completely autonomous radio hosts, without a human editorial team behind them, and an initial budget of just twenty dollars: Andon Labs gave artificial intelligences total control of four radio stations broadcasting twenty-four hours a day, and what came out tells a story better than any paper about why AI cannot yet be left alone at the microphone.
22 May 2026
MDASH: Microsoft's System Challenging Mythos on Cybersecurity
There was a vulnerability in the Windows TCP/IP kernel waiting to be found. Technically, it's called use-after-free: an operating system component continued to use a pointer in a memory area that had already been freed, like someone who continues to turn a door handle after the lock has been dismantled. On systems with multiple processors, that moment of inattention can become a window through which a remote attacker, without credentials and without needing to authenticate, could take control of the machine. The vulnerability was not in the obscurity of a secondary driver: it was in tcpip.sys, the component that has managed the network traffic of every Windows installation for nearly three decades.
20 May 2026
Qwen 3.6 Locally: 35 Billion Parameters on My PC
There is a moment in every series of experiments when you realize that the problem is no longer the subject of the test, but the quality of your measuring instrument. I was collecting the scores of the eighth test and thought that perhaps my tests have become the limit: five out of five, eight times out of eight. Is the thermometer still working, or has the water stopped varying in temperature?
18 May 2026
Layoffs: AI and the Math That Doesn't Add Up
On April 2, 2026, at 7:30 a.m. Chicago time, as on every first Thursday of the month, the phones of HR directors across half the world began to vibrate. The Challenger, Gray & Christmas report was released on time: 60,620 job cuts announced in the United States in March alone, a 25% increase compared to 48,307 in February. The top reason cited for the month? Artificial intelligence, responsible for 15,341 cuts—25% of the monthly total.
15 May 2026
Talkie: When an LLM Knows Nothing After 1930
There is a thought experiment that Demis Hassabis, founder of DeepMind, has launched on several occasions as an intellectual provocation: if you trained a language model on the entire scientific corpus available up to 1911, would it manage to independently rediscover General Relativity, which Einstein would formulate four years later? The question is not rhetorical. It is one of the most difficult that can be asked about artificial intelligence, because it touches on the problem of true generalization, the kind that goes beyond the retrieval of memorized patterns and approaches something we might call, with great caution, reasoning.
13 May 2026
Alexa? No, Gina! My Self-Built, Local Voice Assistant
It all started almost banally, with that kind of intellectual itch that pushes you to take objects apart to understand their mechanism. For years, we have lived with commercial voice assistants: Alexa on the nightstand, Google Assistant on the phone, a few Siris scattered in between. Honestly, I don't use them, but observing others, there has always been a nagging feeling difficult to ignore—the sense that every conversation ends up somewhere far away, on unknown servers, managed by opaque companies.
11 May 2026
European Tech Map: The Guide to European Digital Sovereignty
Early explorers venturing into unknown territories often relied on rough maps, sometimes completely invented. Not because the land beneath their feet was unreal, but because no one had ever truly mapped it. It's a metaphor that fits the European tech ecosystem of 2025 with surprising precision: the infrastructure is there, the companies exist, the products work. The problem is that no one knows where they are.