Smart portal about artificial intelligence and high technology in Kazakhstan
Inference economics, leading open-weight models, global compute metrics, and sovereign breakthroughs.
Inference Economics
Phi-4
16K · Gemini Flash & DeepSeek · up to 1M context
Open Weights Champion
Open weights · Mistral AI
671B MoE · SOTA without vendor lock-in
Global Compute Index
24.8M · 18.3 GW
24.8M chips · KZ Cluster: Nvidia H100 SXM5
Kazakhstan Sovereign Scale
KazLLM 70B · 117 OER
120+ tech companies · KazLLM 70B · 38 OER projects
Live radar: powered by Epoch AI, LMSYS Chatbot Arena, MDDAI registry, and open benchmarks.
Alim Khamitov on the development of the venture industry in Central Asia and new tools for supporting IT entrepreneurs weproject
Где уже сейчас используют казахстанский ИИ Kursiv Media
The most interesting cases were discussed at the large-scale conference AI CRM CONF 2026
Test drive in the rain and a course on artificial intelligence Kapital.kz
The Telecom Set Directorate spoke about quality control, digital labeling of goods, working with domestic manufacturers and the introduction of AI in procurement
I am a city dweller in the classic sense. My background is in law, international law. Before becoming a farmer and livestock breeder, I ran a cosmetics manufacturing plant. I was 30 years old. I began to think about what kind of work I should dedicate my life to. Well, everyone…
Open-source local AI SDK - run AI on-device with no cloud, no API keys. Supports GGUF, RAG, image, music, and video generation, speech-to-text, P2P inference, and more. Cross-platform: Linux, macOS, Windows, Android, iOS.
Chengdu Huitong West-electronic co.,ltd
Chengdu Huitong West-electronic co.,ltd
Chengdu Huitong West-electronic Co., Ltd. is a leading supplier of piezoelectric ceramics, ultrasonic sensors, transducers, piezoelectric actuators, and piezoelectric fans providing reliable solutions for applications in sensing, actuation, and advanced thermal management.
Swipe for more highlights
Flagship open-source projects, MCP tools, multi-agent frameworks, and Kazakhstan datasets.
Catalog and marketplace for Claude Code extensions — 28k+ plugins, 5.8k+ skills, and 6.5k+ MCP servers. Browse skills, subagents, commands, hooks, and plugin collections to extend Claude Code, Claude Desktop, and the Agent SDK.
REA: Reverse engineer anything with AI agents — autonomous binary analysis, decompilation, app behavior inspection down to native binaries via MCP and agentic frameworks.
The agent harness performance optimization system. Skills, instincts, memory, security, and research-first development for Claude Code, Codex, Opencode, Cursor and beyond.
Robots, AI gadgets, and drones — informational cards with links to manufacturer websites.
Browse catalog→Chengdu Huitong West-electronic co.,ltd
Chengdu Huitong West-electronic Co., Ltd. is a leading supplier of piezoelectric ceramics, ultrasonic sensors, transducers, piezoelectric actuators, and piezoelectric fans providing reliable solutions for applications in sensing, actuation, and advanced thermal management.
Centrepolis Accelerator @ Lawrence Tech Univ.
Centrepolis Accelerator at Lawrence Technology University is committed to strengthening supply chain resiliency to drive economic growth and innovation in Michigan. We offer hardtech startups and small to medium-sized manufacturers access to product development services and tailored funding solutions.
Caterpillar Inc.
Caterpillar is committed to helping its customers build a better, more sustainable world. In addition to providing equipment and power solutions to the construction and mining industries, we also deliver advanced technology and autonomy solutions that help keep people safe, optimize operational efficiency and minimize environmental impact
Capsulab
Capsulab is redefining the way wellness and beauty products are created - with our patented digital IoT platform and "Factory in a Capsule" technology, we enable fresh, personalized, and sustainable product manufacturing at the point of use. Our tech key advantages are AI-driven Tech, User Behavior Data & Massive Supply Chain Efficiency
Kazakhstani AI startup Defect AI has been selected for the StartX accelerator at Stanford University GOV.KZ
🔴 Critical PHP vulnerabilities have led to massive shutdowns of websites around the world. What is known now? On October 21, government websites, hospitals and other organizations in several countries temporarily stopped working due to two dangerous vulnerabilities in PHP, the language...
arXiv:2610.28875v1 Announce Type: new Abstract: Large language models increasingly take part in extended conversations where they can both shape and be shaped by another party's views, yet most persuasion research measures only one side of that exchange. We introduce a…
arXiv:2610.28855v1 Announce Type: new Abstract: Progress on foundation model pretraining is gated by who can afford a trial, and relational pretraining is no exception. Recipes are developed inside individual labs, reported under private protocols, and do not accumulate.…
Sovereign language models, national compute cluster, startups, and public sector AI.
The Telecom Set Directorate spoke about quality control, digital labeling of goods, working with domestic manufacturers and the introduction of AI in procurement
The most interesting cases were discussed at the large-scale conference AI CRM CONF 2026
I am a city dweller in the classic sense. My background is in law, international law. Before becoming a farmer and livestock breeder, I ran a cosmetics manufacturing plant. I was 30 years old. I began to think about what kind of work I should dedicate my life to. Well, everyone…
💙 Social Impact: 5 features on the astanahub.com platform that you can use today! We have collected 5 Social Impact activities available to participants of the Astana Hub ecosystem: 📝 Share your experience — about your participation in AI & Digital Bridge 2026 on the astanahub.com platform…
Serious vulnerabilities have been discovered in the PHP programming language, leading to a critical cybersecurity update. World governments and organizations are calling for their sites to be shut down if necessary fixes cannot be applied in a timely manner. Vulnerabilities identified using...
⚖️ Anthropic предложила блокировать ИИ-модели, если они могут привести к гибели 50 человек По данным Politico, компания подготовила законопроект, который позволит через суд останавливать выпуск потенциально опасных нейросетей. Основанием станет непосредственная угроза гибели…
В Атырауской области школьники осваивают робототехнику благодаря малому гранту Ratel.kz
Новости по теме Искусственный интеллект TengriNews
Любой каприз за вашу подписку. В Казахстане не жалеют денег на обучение детей через ИИ TengriNews
💙 Social Impact: 5 features on the astanahub.com platform that you can use today! We have collected 5 Social Impact activities available to participants of the Astana Hub ecosystem: 📝 Share your experience — about your participation in AI & Digital Bridge 2026 on the astanahub.com platform…
Стартап казахстанца достиг $1 млн ARR за 10 месяцев. Call2Action.ai использует ИИ для найма сотрудников в США и нацеливается на $10 млн.
Stepful (YC S21) trains people for healthcare jobs, combining AI-supported learning with local hands-on training. The company trained 30,000 people last year and now works with 39 health systems, including seven of the 10 largest in the U.S. It recently raised a $55 million…
Интересно, а когда роботы от Unitree сдадут экзамены на дан по карате? Когда их маркетологи возьмутся за этот курс продвижения?:) Сегодня Unitree выпустили этот ролик с новым роботом - телохранителем и меня (Ани Асланян, основатель @blockchainrf) впечатлила демонстрация базовых…
Голосовой помощник Gemini теперь доступен пользователям Apple CarPlay в США.
Стартап Recursive представил модель Flint, демонстрирующую рекурсивное самоулучшение ИИ.
arXiv:2610.29057v1 Announce Type: new Abstract: Users increasingly rely on large language models (LLMs) to analyse collections of independent inputs, such as product reviews or social media posts, and return a single aggregated answer. This setting requires multi-instance…
NVIDIA вместе с Нобелевским лауреатом выпустили ИИ-модель для дизайна биомолекул, которая в вычислительных тестах обходит конкурентов по доле успеха и скорости Команда Дэвида Бейкера и Nvidia опубликовали веса и код своей ИИ-модели RFD4-Proteina. Модель работает в общем…
Mistral и NVIDIA развивают ИИ для инженерных процессов, ускоряя симуляции.
NVIDIA разработала RFD4-Proteina для проектирования белков с помощью ИИ.
If you have two or three Macs at home and want to run a large AI model locally, the memory on one computer may not be enough. exo allows you to combine multiple devices into one AI cluster. A model that does not fit into the memory of one computer can be distributed among several...
🌟 CLM-8B: an open and fast model for selecting agent actions Stanford, in collaboration with NVIDIA, has made the first model of the Contrastive Language Models class publicly available. This is an open analogue of Jev, which was presented on September 15 by the TypeSafe AI startup of a former researcher...
✔️ Dream Labs presented a model for controlling unfamiliar robots Four former NVIDIA researchers from the GR00T robotics project team came out of stealth with their startup and opened the public API of the Dream-1 model. The company was led by Joel Jang, who led the GR00T…
arXiv:2610.28826v1 Announce Type: new Abstract: Automated judges of clinical artificial intelligence (AI) outputs are necessary, but large language model (LLM) judges are costly and slow at scale. We compared Jev, a decision language model that returns structured judgments, with…
Исследование Google Research демонстрирует, как ИИ может улучшить материнский уход.
A rush of new data centers could bring air pollution roaring back to levels not seen since the 2000s in Texas, which has become ground zero for the physical infrastructure behind generative AI. New and proposed data centers are demanding a tremendous amount of electricity. How…
Tesla's revenue was flat and its expenses skyrocketed as it pressed the accelerator on its Optimus robots, AI, and autonomous driving.
The news comes as Agility inches closer to its IPO via SPAC and after it announced the latest version of Digit. | Source: Agility Robotics AGCO and Agility Robotics have signed a commercial agreement to deploy Agility’s Digit humanoid in AGCO’s manufacturing facility. The U.S.…
Open-source local AI SDK - run AI on-device with no cloud, no API keys. Supports GGUF, RAG, image, music, and video generation, speech-to-text, P2P inference, and more. Cross-platform: Linux, macOS, Windows, Android, iOS.
Video generation models can now depict remarkably realistic motion. Yet it is unclear how to use this motion beyond pixels: to animate a specific 3D asset and replay the animation in its 3D scene. For instance, a video model can show a monkey jumping over a crossbar, but it cannot return that jump as replayable 3D motion placed in the 3D scene. Here we present a model that directly co-generates the video and 3D motion for any supplied rig, along with a camera that registers this motion to the scene. Prior video-motion models either handle only human skeletons, often adding billions of paramete
Просишь ИИ сделать анимацию для вступления к видео. В голове уже есть представление о том, как она должна выглядеть, но не знаешь, как называется нужный эффект. Остаётся писать что-то вроде «сделай покруче и посолиднее», а результат зависит от удачи.motion-prompt создан именно для решения этой проблемы. Это своего рода словарь визуальных эффектов для работы с ИИ. Для каждого эффекта предусмотрены анимированный пример, краткое объяснение и готовая формулировка запроса.Сейчас в словаре 120 эффекто
Agentic artificial intelligence (AI) systems hold immense potential for automating complex workflows in materials discovery and design, but extending them to domain-intensive mechanistic investigations remains challenging. Here, we present GBClaw, an agentic AI framework for investigating the structure and deformation of general grain boundaries (GBs). Although general GBs constitute the vast majority of real-world interfaces in polycrystalline materials, they remain understudied because their structural complexity and interfacial mismatch complicate experimental and modeling research. GBClaw
Когда Codex работает с десктопными приложениями вроде Blender, ему приходится делать скриншоты, анализировать интерфейс и управлять мышью. При этом все окна находятся на основном экране, из-за чего пользователь и ИИ-агент мешают друг другу.yomado — инструмент для macOS, который создаёт виртуальные дисплеи и позволяет выделить ИИ-агенту отдельное рабочее пространство, не занимая физический экран. Нужные приложения можно перенести туда, чтобы агент работал с ними независимо от пользователя.Можно с
arXiv:2610.29023v1 Announce Type: new Abstract: Large language model (LLM) agents are often improved by iterating until a fixed verifier is done. Once that verifier saturates, the loop has nothing left to climb. We instead score agents against the current population, and a…
arXiv:2610.29046v1 Announce Type: new Abstract: Large language models (LLMs) have shown considerable promise in synthesizing formal specifications from source code. However, existing approaches primarily focus on functional contracts and loop invariants, while largely neglecting…
arXiv:2610.29057v1 Announce Type: new Abstract: Users increasingly rely on large language models (LLMs) to analyse collections of independent inputs, such as product reviews or social media posts, and return a single aggregated answer. This setting requires multi-instance…
Стартап казахстанца достиг $1 млн ARR за 10 месяцев. Call2Action.ai использует ИИ для найма сотрудников в США и нацеливается на $10 млн.
Интересно, а когда роботы от Unitree сдадут экзамены на дан по карате? Когда их маркетологи возьмутся за этот курс продвижения?:) Сегодня Unitree выпустили этот ролик с новым роботом - телохранителем и меня (Ани Асланян, основатель @blockchainrf) впечатлила демонстрация базовых…
😎 Avito again participates in the hh.ru Employer Rating! Last year, thanks to your support, we entered the top 3 IT companies in the country. If what we do resonates with you, vote for Avito this year too ❤️ How to support the company: - Go to hh.ru and log in...
🌟 CLM-8B: an open and fast model for selecting agent actions Stanford, in collaboration with NVIDIA, has made the first model of the Contrastive Language Models class publicly available. This is an open analogue of Jev, which was presented on September 15 by the TypeSafe AI startup of a former researcher...
🤖 Anthropic released Claude Duet: a personal agent who writes to you first. Claude already had agents, but they had to be launched and waited. Duet works differently: you throw a task into a messenger-like thread, it does it in the background on your cloud computer and writes when...