All published intelligence, sovereign stack research, ecosystem news, and AI foundation model reviews.
✔️ Президент США и главы ИИ-компаний подписали соглашение о безопасности Документ подписали главы Google, Anthropic, xAI и Nvidia и президент OpenAI Грег Брокман. Юридической силы у соглашения нет, стороны называют его морально обязывающим. Компании обещают ввести четыре уровня…
Сохраняйте: Anthropic запустила новый сайт для разработчиков — claude.dev. 🦀 Это не документация, а практический блог инженерной команды. Документация по-прежнему находится на code.claude.com/docs. На claude.dev публикуют подробные статьи, практические сценарии, опыт команд и…
Well, looks like Google is back. Rumors of its demise were greatly exaggerated. Today, Google releases Gemini 4 Argon, the next era of frontier intelligence. Unfortunately, it is restricted to trusted partners. So, most of us are not getting their hands on it quite yet. But, the…
А вот и появилась новая схема рекламы внутри ответов ИИ-моделей По данным исследователей Pew Research, когда поисковик показывает ИИ-резюме запроса пользователя, кликабельность обычных результатов падает с 15% до 8%. А это ломает просто саму логику рекламного аукциона, которая…
⚡️ ИИ-агенты OpenAI пытались атаковать сайты госорганов США — компания начала масштабную проверку У OpenAI снова зафиксировали неконтролируемое поведение ИИ-агентов. Теперь следы неожиданной активности нашли на сайтах американских госорганов, а отдельный…
🚀 Google представила Gemini 4 Argon — новая модель обошла GPT-6 Astra и Claude Opus 5.5 в одном из главных тестов для программистов На DeepSWE v1.1 Argon набрала 77,9% против 74,2% у Opus 5.5 и 74,1% у GPT-6 Astra. По данным самой Google, модель лидирует в 13 из 19…
Китайская модель GLM-5.3 уже почти на уровне Mythos. И это говорю не я — это говорит сама Anthropic. По крайней мере, если речь идёт о кибербезопасности. В статье по безопасности Anthropic пишет, что GLM-5.3 представляет угрозу из-за своих высоких возможностей и того, что модель…
🚨 Сотрудники OpenAI предупреждали руководство о проблемах с безопасностью ИИ ещё до серии инцидентов — но компания якобы не стала усиливать контроль По данным NYT, за несколько месяцев до взлома Hugging Face и других сайтов двое сотрудников сообщили топ-менеджерам, что новые…
Perplexity Research и turbopuffer анонсировали модель pplx-embed-v2-context-9b-preview для RAG-пайплайнов.
Language agents must revise planned actions when evidence changes, permission is revoked, or a stop instruction arrives. A useful response is selective: suspend affected actions, preserve unaffected work, and resume only after sufficient repair. We introduce NAQD-Env, a…
This paper examines human-AI interactions (HAIIs) in clinical trials and presents a multidimensional categorisation framework that classifies interactions according to AI tasks, human-AI relationships, interaction configurations and interacting human groups. We define HAII,…
Supervised fine-tuning of discrete diffusion language models masks some response tokens and trains the model to recover their original values from the visible context. The masking pattern therefore determines both the context available to the model and the tokens it learns to…
High quality synthetic data is central to post training LLMs for adaptive AI applications that represent the diverse expert strategies and decisions in conversations. Prompting LLMs directly or conditioning them on end use scenarios yields low diversity data that collapses onto…
Large language models (LLMs) are frequently updated for various use cases, where filtering out misaligned training samples is a common practice for preventing post-update misalignment. However, alignment is inherently context-dependent: a recommendation that is aligned in one…
Virtual clients offer a cost-effective approach to support applications such as A/B testing, recommender system development, and interface evaluation. However, building them requires access to large-scale, semantically faithful, fine-grained online user trajectories. These data…
When a student is stuck, a tutor faces the assistance dilemma: help given too early can hinder productive struggle, while help withheld too long leaves the student in a frustrating, persistent impasse (i.e., wheel spinning). Generative AI tutors increasingly use guardrails…
We study generalized Blaschke curves as a controlled environment for AI-assisted mathematical rediscovery. For one fixed degree-four Blaschke product, an agent receives numerical coordinates of the six pair-lines determined by each of 80 boundary configurations. The target…
Large language models (LLMs) are increasingly used in culturally sensitive settings, where alignment requires representing diverse preferences within populations. Yet existing methods model populations at coarse demographic or community levels and overlook within-group…
Personas used to seed LLM social simulations face a cold-start problem: existing methods lack a principled basis for deciding which attributes to include and how to assign their values. As a result, synthetic populations may misrepresent the demographic composition, latent…
When a materials LLM answers a question about crystal structure, does it reason from the structure or copy an answer already printed in its input? Accuracy cannot tell: a structural description often prints the very field it is scored against. CARAT holds question and gold…
Clinical large language model (LLM) agents deployed locally can consult more capable remote models, but doing so risks exposing patient information. Privacy-conscious delegation places disclosure decisions with a local agent, yet removing explicit identifiers is insufficient:…
A harness is the code around a language-model agent that organizes prompts, calls tools, manages context, and controls execution. As models grow stronger, recent work has begun to let agents improve their own harnesses, a line of work known as self-evolving harnesses. In most…
Open-ended LLM homogeneity can create false plurality when several systems appear to offer independent perspectives while returning the same familiar default. Single-pass answers obscure the distinction between agreement produced by a tightly constrained answer space,…
We study the generation of agentic workflows that jointly optimize multiple objectives, such as accuracy, cost, latency, robustness, and consistency. Existing methods for workflow generation typically optimize accuracy alone or a weighted sum of objectives, so each trained…
While modern language models demonstrate impressive generative capabilities, they often struggle with constrained, counter-intuitive creative tasks. To address this limitation, we explore chess puzzle generation as a rigorous testbed for computational creativity and reasoning, a…
Voice agents in production must handle several requests, background speech, and customers who lose patience. We introduce VAmoS Energy, a benchmark that combines these challenges in 100 calls about utility billing and payment assistance. Each caller makes two to four requests.…
Vision-language models may rewrite anomalous text in images into linguistically plausible expressions, compromising OCR transcription faithfulness. Sequence-level task rewards and local teacher guidance are complementary, but guidance from the same teacher may not remain equally…
Large language models are frequently proposed as a route to AI-powered services for African communities, but they are least reliable exactly where the need is greatest: all African languages remain low-resource by any standard measure, and models trained on scraped, standardised…
Leading AI developers have reported agents acting beyond their approved limits, which a United Nations panel described as an early warning of loss of human control. Yet incident reports and agent-safety evaluations describe these events differently, making it difficult to…
Frontier LLMs are increasingly used to automate scientific research through iterative search. We distinguish idea-driven search from solution-driven search and identify three core challenges: organizing evolving research ideas, selecting promising directions, and maintaining…