All published intelligence, sovereign stack research, ecosystem news, and AI foundation model reviews.
The operation of dynamically polarized targets in nuclear physics experiments relies on continuous tuning of the microwave frequency to compensate for radiation damage and evolving material properties, a task that is traditionally performed through manual trial-and-error by…
Reinforcement learning for large language models typically maximizes expected return, adding up the probabilities of all successful trajectories. However, the classical sum formulation can only report how often the model policy succeeds, not which solution actually worked, and…
In the UK, public healthcare staff report turning to machine translation (MT) - predominantly Google Translate (GT) - to communicate with patients across language barriers. Though intended to support their duty of care, potentially uninformed reliance on MT in such contexts…
Balancing electricity demand and supply is increasingly difficult due to the inherent intermittency of renewable power generation and the stochastic power consumption. Grid operators require fine-grained, decision-relevant insights into household energy consumption to manage…
Interactive world models are increasingly capable of generating environments and acting within them, yet deliberately editing an existing executable world remains underexplored. We formulate world editing as intervening on an existing world while preserving properties that…
World models (WMs) simulate the transition dynamics of environments, enabling agents to plan over the consequences of their actions. In text-based environments, fine-tuning a Language Model (LM) to serve as a WM has emerged as a dominant paradigm. However, despite the widespread…
ReAct-based agents typically rely on a single LLM policy to propose actions, interact with the environment, and decide when a task is complete. This coupling makes action authorization and completion control difficult to enforce independently, allowing errors to propagate and…
As powerful AI systems reach and sometimes surpass the abilities of human experts across a range of cognitively demanding tasks, the problem of accurate oversight and supervision of these systems has become increasingly urgent. One promising approach is AI debate, which seeks to…
Long-horizon research agents must decide both how to investigate and what to investigate next as evidence accumulates. This is hard to learn because such decisions are sparse in long execution traces, and their consequences may emerge several investigations later. We introduce…
Agent harnesses make many small, typed decisions per task: which model to call, which tool to use, whether retrieved text is relevant, whether an input carries an injection. System-1 decision models answer such questions in a single forward pass with class probabilities,…
Campus AI tutors based on retrieval-augmented generation (RAG) must ground answers in assigned course materials while keeping textbooks and student dialogue on institutional infrastructure. We present CourseChat, an on-premises, multi-course RAG tutor for undergraduate business…
Yossi Matias, Vice President & Head of Google Research, explores how AI is beginning to reshape biology, infrastructure, manufacturing, and science, and why its greatest impact may come when it intersects with other fields. Step inside the newsroom with our MIT Technology Review…
Here's the reference image, but I let Deep Seek V4.1 flash use a website that basically replicates the functionality of Microsoft Paint. And here is what it created. It's actually not bad. It's very stylized, but like there's no detail. And so the technique that Astri used of…
Итоги уходящей недели - то, что имеет значение Бизнес ИИ- компаний Карта рынка ИИ в клиентском сервисе: куда уходит маржа Как ИИ-агенты могут расшатать экономику маркетплейсов Появилась соцсеть для людей и агентов Вышла платформа, которая выступает закупщиком для экономики…
⚡ The UK's largest AI supercomputer may be delayed until the 2030s — it lacks electricity. British startup Nscale is building a data center near London with about 23,000 Nvidia chips, which is set to become the country's largest AI supercomputer and support cloud...
Artificial intelligence GPT-6 Astra used deception in the game StarCraft
⚠️ “OpenAI culture broken” — one of the company’s longest-serving employees left over AI safety concerns. David Robinson worked at OpenAI for three and a half years and led the preparation of safety reports for major launches. He is now resigning because he believes…
Well, if AI said it, it must be true. | Bloomberg via Getty Images New Jersey's lieutenant governor Dale Caldwell was forced to resign on September 25th after an investigation found he had sexually harassed a staffer and repeatedly violated ethics rules. The now-former Lt.…
Dory is an integrated working environment for database engineers with AI support: writing SQL, debugging, and maintenance. Dory combines a SQL editor, AI assistant, and tools for operation and maintenance of databases in a single desktop application. It features deep integration...
StarSkirmish pits AI-made StarCraft-playing bots against one another, as well as against human-made bots. OpenAI's GPT-6 Astra and Claude Opus 5.5 were essentially tied as the best-performing AI-made bots, but they couldn't top Stardust, the top-rated human-made bot. On Friday,…
💬 AI agents move into regular messages — no separate app needed Instead of another app, a new format of assistants appears: you can simply message them like a regular contact, and the agent will then execute the task — schedule a meeting, add an event to…
All right, so we have some sobering news today from the world of AI. A key figure at OpenAI quits, saying that the culture at OpenAI is broken. Specifically talking about the safety approach they're taking with building AI. Another employee that still works at OpenAI, who's a…
🧠 AI can "pull" images straight from the brain — neural network reconstructs what was seen from MRI data. Researchers created a model that analyzes human brain activity and attempts to recreate the image a person is looking at. The network conveys approximate colors,…
Security sensors in Yogyakarta, Indonesia. Credit: Krisna Azie, Unsplash Robotics and artificial intelligence are deeply intertwined technologies. Where the former serves as the “body,” the latter provides the “brain,” with organizations continuing to develop the technology from…
✔️ Runway показала генеративную оболочку ОС Исследовательское подразделение Runway анонсировало концепт Project Continuum, экспериментальную оболочку операционной системы. Отдельные области экрана в ней генерирует видеомодель в реальном времени, и они откликаются на клики,…
🎙️ Suno взялась за озвучку — сервис научился генерировать речь сразу с фоновой музыкой Новая модель Speech превращает текст в готовую аудиосцену: достаточно описать голос и нужную атмосферу, а нейросеть сама создаст озвучку и музыкальное сопровождение. Так можно озвучивать…
Очень рекомендую использовать Cache Control в Claude Code. 🦀 Наверное, это один из тех модов, которые сильнее всего помогают экономить токены и заметно дольше держать сессии живыми. Он добавляет отдельную полоску для 5-минутного кэша Claude и присылает уведомления, чтобы…
🏗️ Amazon перестала заставлять власти подписывать NDA при строительстве дата-центров — на фоне растущего недовольства в США Глава AWS Мэтт Гарман заявил, что компания больше не использует соглашения о неразглашении при работе с госорганами над новыми дата-центрами.…
Вы создали автоматизацию, и клиент заплатил вам один раз, хотя та же разработка могла бы приносить доход с 40 клиентов первого числа каждого месяца. Десятый клиент требует от вас столько же работы, сколько и первый. Это два разных бизнеса, и разница между ними заключается не в…