All published intelligence, sovereign stack research, ecosystem news, and AI foundation model reviews.
Генеральный директор Anthropic Дарио Амодеи призывает замедлить развитие ИИ, акцентируя внимание на необходимости оценки безопасности и координации между странами. В противовес этому, генеральный директор Nvidia Дженсен Хуанг выступает за быстрое развитие, отвергая опасения по…
Статья описывает растущую тенденцию неудач стартапов и проектов в области ИИ, которые не могут выжить в условиях конкуренции с крупными компаниями, такими как OpenAI и Google. Многие инициативы закрываются из-за недостаточного финансирования, технических проблем или низкого…
В браузере запустили полноценного агента для программирования на модели всего с 2 млрд параметров..
Listen to the session or watch below Employees at the world’s leading AI labs are saying there’s a real possibility that advanced AI could destroy humanity. Are they right? Or is this more scaremongering and hype? Watch a conversation unpacking AI extinction fears: where they…
OpenAI, Anthropic и Google сотрудничают по вопросам безопасности ИИ в течение нескольких недель. Это сотрудничество стало результатом призыва CEO Anthropic к объединению отрасли для замедления разработки передового ИИ и снижения катастрофических рисков. Обсуждаются стандарты…
😈 “AI can kill us all”: how fear around AI reached a new level in a week Over the past few days, the discussion about the safety of AI has sharply expanded beyond the expert community. Time even put the theme on its cover. Researchers from leading AI laboratories began publicly...
Основатели компании Artificial Intelligence Underwriting Company (AIUC) Рун Квист и Раджив Даттани разрабатывают решение для управления рисками, связанными с агентами ИИ в компаниях. Они создали стандарт AIUC-1 для аудита и сертификации агентов ИИ, что обеспечивает соответствие…
👀 Project Lily: OpenAI contractors read real conversations of ChatGPT users OpenAI is hiring hundreds of contractors to check real conversations of ChatGPT users as part of a project codenamed Project Lily, 404 Media has found out. 🤔 Reviewers read requests...
This is today’s edition of The Download, our weekday newsletter that provides a daily dose of what’s going on in the world of technology. The AI industry has taken a doomer turn. What now? AI chiefs Dario Amodei, Sam Altman, Elon Musk, and Demis Hassabis are suddenly all in…
Модели ИИ нуждаются в большем количестве данных о биологии, и OpenAI платит за их создание.
What happens if you completely entrust the store to artificial intelligence? An SFGATE journalist checked out the Andon Market experiment in San Francisco.
У Claude Code начинает появляться собственная экосистема модов.
President of Kazakhstan Kassym-Jomart Tokayev and Prime Minister of South Korea Han Sung-sook got acquainted with the Alatau Smart City projects. They were shown an air taxi control system created with the participation of Korean companies. During the presentation the main…
Images can be watermarked for copy protection. Now, get this. Claude AI announced that they are watermarking the text you generate with it. From when? When does this start? Well, Anthropic is rolling this out right now. Yep. I am not talking about this because I agree with it,…
Казахстан планирует перейти от наращивания майнинговых мощностей к созданию полноценной индустрии цифрового майнинга. В нее войдут криптобиржа, инструменты для работы с цифровыми активами, токенизация, хранение активов и платежные решения, передает Tech-news. Такое поручение на…
Archives of Kazakhstan and Korea agreed on cooperation in the field of AI and digitalization Kazinform
The article raises questions about the need for AI security measures amid recent events, including the resignation of a researcher from Anthropic and growing concerns about uncontrolled progress in AI. Anthropic CEO Dario Amodei proposes a three-step plan...
Инженер DeepSeek заявил, что не хочет, чтобы Anthropic получила контроль над AGI..
Зачем Казахстану ученые, способные работать на стыке науки, ИИ и автоматизации - forbes.kz.
Industrial maintenance systems involve multiple interacting assets and shared resources, making it challenging to balance reliability and operational cost using a single decision framework. While recent work has focused on reinforcement learning (RL) for maintenance scheduling,…
The increasing deployment of AI agents in long-horizon tasks yields massive execution logs. Diagnosing failures within these records is crucial for reliability, as it transforms outcome-level signals into actionable interventions. The sheer scale of the data renders human review…
Safety research often focuses on model-generated harms, but users may also direct hostility, coercion, and adversarial pressure at models. Understanding how and when that occurs is essential for accurately interpreting model behaviour, alignment drift, and real-world deployment…
LLM judges are increasingly used to evaluate and improve AI-generated outputs, yet their reliability for complex professional work remains unclear. We study this problem through Vibe Patenting, an end-to-end patent-drafting testbed for AI-agent evaluation. A separately-invoked…
Web agents can utilize reusable tools to reduce the cost and latency of low-level browser interaction, but automatically discovered tool collections can be large, redundant, and poorly aligned with user demand. We present AutoTailor, a meta-agentic framework for constructing and…
Timeseries multimodal large language models (TS-MLLMs) have recently begun leveraging the reasoning capabilities of large language models (LLMs) for question-answering tasks. However, these models often fail to capture dynamic temporal patterns, providing only implicit reasoning…
Voice agents often need to collect names, addresses, identifiers, dates, and times exactly, yet end-to-end benchmarks obscure where capture fails. We introduce $τ$-Elicitation, a 200-task voice benchmark spanning 10 entity types, controlled difficulty, caller realisms, and three…
Chemotherapy improves survival for some patients with breast cancer, but doctors cannot reliably predict who. Current guidelines rely on recurrence scores as a proxy for treatment benefit, which may contribute to the overprescription of chemotherapy. Here we present a causal…
Large Language Model (LLM) agents offer a promising path toward autonomously managing long-term physical tasks without human intervention. However, physical tasks require agents to continuously observe the environment, make consequential actions, and remain effective as the…
Persistent agents need evaluations that distinguish identity facts they can recall from those they express and enact. We introduce PAI-Bench, a provider-neutral benchmark for fidelity to a versioned, update-governed identity contract. It separates recall, composition, behavioral…