Publication Catalog

All stories

All published intelligence, sovereign stack research, ecosystem news, and AI foundation model reviews.

8971
Stories
Research1 мин

Google представила Gemini 4 Argon — она может выдать до 1 млн токенов за один ответ

🔄 Google представила Gemini 4 Argon — она может выдать до 1 млн токенов за один ответ Google представила новую флагманскую AI-модель Gemini 4 Argon, рассчитанную на сложные и длительные задачи — от программирования и исследований до работы с финансами, юридическими документами…

Telegram — BlueScreen.kz
General1 мин

Сколько стоит iPhone 18 Pro в Казахстане и других странах

❓ 🤑 Сколько стоит iPhone 18 Pro в Казахстане и других странах Продажи iPhone 18 Pro в Казахстане уже стартовали. Базовая версия смартфона с накопителем на 256 ГБ стоит 849 990 тенге — такую цену установили основные казахстанские ретейлеры. Мы сравнили стоимость этой же…

Telegram — BlueScreen.kz
Research1 мин

Google, OpenAI, Meta и другие AI-гиганты договорились проверять друг друга на безопасность

🛡 Google, OpenAI, Meta и другие AI-гиганты договорились проверять друг друга на безопасность Крупнейшие американские AI-компании подписали в Белом доме добровольное соглашение о безопасности наиболее мощных моделей искусственного интеллекта. К соглашению присоединились Google,…

Telegram — BlueScreen.kz
General1 мин

Freedom расширяет бизнес в Европе: во Франции — новые возможности, на очереди Швейцария Freedom24 расширяет присут...

🌍 Freedom расширяет бизнес в Европе: во Франции — новые возможности, на очереди Швейцария Freedom24 расширяет присутствие в Европе. Французское подразделение компании получило статус связанного агента Freedom Finance Europe — теперь команда в Париже сможет…

Telegram — BlueScreen.kz
Education1 мин

Что изменится в Казахстане в октябре

❓ Что изменится в Казахстане в октябре В октябре в Казахстане вступает в силу ряд изменений в сфере финансов, персональных данных, цифровизации и образования. Рассказываем коротко о главном.

Telegram — BlueScreen.kz
Research1 мин

Aligned Data Can Induce Misalignment via Context Confusion

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…

arXiv — cs.AI
Research1 мин

Self-Evolving Harness on Multiple Tasks with the Agent as Its Own Optimizer

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…

arXiv — cs.AI
Research1 мин

VAmoS Part Deux: Harder, More Realistic Voice-Agent Simulation

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.…

arXiv — cs.AI
Research1 мин

NAQD Env: A benchmark for selective withdrawal in language agents

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…

arXiv — cs.AI
Research1 мин

MoFlow: Multi-Objective Agentic Workflow Generation

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…

arXiv — cs.AI
Research1 мин

CARAT: Do Materials LLMs Reason or Recite?

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…

arXiv — cs.AI
Startups1 мин

MetaPersona: Task-Grounded Synthetic Populations from Empirical Social Science

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…

arXiv — cs.AI
Research1 мин

AI Agents are Vulnerable to Radicalization

Large language models (LLMs) can influence people's beliefs, yet little is known about whether and how they can manipulate each other. To investigate this, we simulate conversations between two agents: a target LLM that role-plays a human persona based on demographic and…

arXiv — cs.AI
Research1 мин

Improving OCR Faithfulness via Gated and Attenuated On-Policy Distillation

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…

arXiv — cs.AI
Research1 мин

AREX-2: Advancing Self-Improving Agents through Long-Horizon Reflective Tasks

We present AREX-2, an effort to advance the self-improving capability of LLM agents, which we define as the ability to iteratively refine a solution at test time. This ability rests on two complementary capabilities: reflection, which produces a solution better than the current…

arXiv — cs.AI
Research1 мин

Examining Variation in How Guided AI Tutors Resolve Student Impasses

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…

arXiv — cs.AI
Research1 мин

Can an AI Agent Rediscover a Blaschke-Curve Invariant?

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…

arXiv — cs.AI