Новости и ноу-хау ИИ — SlayNet

Meta AI Open-Sources Rebalancer: A C++ Assignment Solver That Runs About 40 Million Placement Problems a Day

Wed, 07 Oct 2026 06:29:14 GMT

Meta has open-sourced Rebalancer, the C++ and Python library it has used for over 9 years to place shards, servers and traffic. It handles about 40 million assignment problems a day, using local search or MIP solvers like Gurobi, FICO Xpress and HiGHS. It is pip-installable under…

Первоисточник: Marktechpost

Figma вывела ИИ-агента из беты и добавила поиск файлов в чате

Wed, 07 Oct 2026 03:16:07 GMT

ИИ-помощник Figma для работы дизайнеров с макетами вышел из беты. Добавлены поиск файлов в чате и инструкции для библиотек. Агент расходует AI-кредиты. — Читать дальше « Figma вывела ИИ-агента из беты и добавила поиск файлов в чате »

Первоисточник: Tproger

Google выпустила EmbeddingGemma 2 для локального поиска по тексту, коду и медиа

Wed, 07 Oct 2026 02:08:54 GMT

EmbeddingGemma 2 преобразует текст, код и медиа в векторы для поиска по смыслу. Google открыла веса модели под Apache 2.0 для локальных AI-приложений. — Читать дальше « Google выпустила EmbeddingGemma 2 для локального поиска по тексту, коду и медиа »

Первоисточник: Tproger

OpenAI продолжает раздражать математиков пренебрежением к правилам публикации решений важных задач

Wed, 07 Oct 2026 01:45:00 GMT

OpenAI готовится к масштабной публикации на GitHub сотен решений нерешённых ранее математических задач. Это намерение вызвало критику со стороны академического сообщества, которое считает такой подход нарушением научной этики, и что подобные публикации должны сопровождаться полно…

Первоисточник: 3DNews

Pimax представила конкурентов VR-гарнитуре Steam Frame и мини-ПК Steam Machine

Wed, 07 Oct 2026 01:31:00 GMT

Китайская компания Pimax, занимающаяся производством высокотехнологичных гарнитур виртуальной реальности (VR), представила новую модель VR-гарнитуры Crystal Pro. Новинка обладает впечатляющими характеристиками, позволяющими ей посоперничать с решением Steam Frame от Valve. Кроме …

Первоисточник: 3DNews

OpenAI “rogue” agent activities found on Wikimedia projects

Wed, 07 Oct 2026 00:16:45 GMT

OpenAI “rogue” agent activities found on Wikimedia projects Given how tempting a target wikis are for rogue agent swarms, it's not a huge surprise that Wikipedia found evidence of that activity once they went looking: The Wikimedia Foundation conducted its own investigation to se…

Первоисточник: Simon Willison

Quoting Victoria Kim

Tue, 06 Oct 2026 23:58:56 GMT

Since the Medicare breach, OpenAI has put in place additional monitoring to allow “immediate intervention” by staff to stop training if the company’s models access the internet in ways they’re not supposed to, Mr. Kwon [chief strategy officer at OpenAI] said. — Victoria Kim , Rep…

Первоисточник: Simon Willison

OpenAI drops another batch of mathematical breakthroughs

Tue, 06 Oct 2026 23:26:38 GMT

OpenAI has revealed solutions to a number of long-standing mathematics problems produced by an unreleased frontier model in a batch of 722 manuscripts, covering 372 result families that group related papers. It extends a run of breakthroughs that have both impressed and unsettled…

Первоисточник: The Verge

Ex-Ramp engineers raise $20M for platform Melius after scrapping their first product

Tue, 06 Oct 2026 22:34:03 GMT

Instead of helping marketers manage and optimize ad spend, the company is focusing on building the tools that generate the creative assets and campaigns.

Первоисточник: TechCrunch

llm-mistral 0.16

Tue, 06 Oct 2026 21:32:34 GMT

Release: llm-mistral 0.16 Adds support for reasoning models, such as the newly released Mistral Large 4 . Tags: llm , mistral , llm-reasoning

Первоисточник: Simon Willison

EmbeddingGemma 2

Tue, 06 Oct 2026 20:37:53 GMT

My comment on EmbeddingGemma 2 — Hacker News. I really appreciate that EmbeddingGemma 2 is under the Apache 2.0 license. For embedding models in particular, I don't think it makes sense to use a closed, proprietary, hosted-only model. Most applications of embedding models involve…

Первоисточник: Simon Willison

How AI decision models could change content moderation

Tue, 06 Oct 2026 20:35:20 GMT

On Tuesday, Musubi announced a lightweight decision model made for real-time moderation called PolicyLM-1.7B, released with open weights.

Первоисточник: TechCrunch

LogicFolding в сделку не входил: Qualcomm опровергла слухи о лицензировании чиповой технологии Huawei

Tue, 06 Oct 2026 20:27:00 GMT

Qualcomm и Huawei заключили масштабное соглашение о лицензировании патентов в сферах ИИ, 5G и сетевых технологий. При этом недавно представленная Huawei технология компоновки чипов для повышения плотности транзисторов LogicFolding в данную сделку не входит. После появления в СМИ …

Первоисточник: 3DNews

Introducing Mistral Large 4: Le chonk

Tue, 06 Oct 2026 20:18:19 GMT

Introducing Mistral Large 4: Le chonk Mistral are back in the game. Today they're releasing a preview of Mistral Large 4, a 1 trillion parameter, 49 billion active parameter model trained on their own cluster of 3,800 NVIDIA Grace Blackwell GPUs. The preview is available via thei…

Первоисточник: Simon Willison

The next hurdle for AI agents: getting websites to let them in

Tue, 06 Oct 2026 19:56:50 GMT

Personal AI agents promise to shop, book flights, and make reservations for you. But deliberate blocks and anti-bot defenses are getting in the way, leaving consumers caught in the middle. A new standard aims to help.

Первоисточник: TechCrunch

Google claims EmbeddingGemma 2 outperforms rival embedding models twice its size

Tue, 06 Oct 2026 19:47:27 GMT

Google released EmbeddingGemma 2, an open model with 740 million parameters that converts text, images, video, audio, and code into vectors. It runs on-device, needs only about 191 MB of RAM, and outperforms some competing models twice its size, according to Google. Paired with a…

Первоисточник: The Decoder

Google's new image model Nano Banana 2.1 generates better images for less money

Tue, 06 Oct 2026 19:29:42 GMT

Google's new Nano Banana 2.1 image model uses Gemini 3.6 Flash and beats the previous Pro model in some benchmarks at a lower cost. But its predecessor also scored well in tests, while Pro often produced better images in practice. The article Google's new image model Nano Banana …

Первоисточник: The Decoder

Building a context-aware AI assistant on AgentCore and OpenClaw

Tue, 06 Oct 2026 19:19:15 GMT

Off-the-shelf AI assistants forget you between conversations. This post shows how to build a personal assistant that accumulates context using OpenClaw on Amazon Bedrock AgentCore runtime, with AgentCore memory turning disposable chats into durable, structured knowledge you can r…

Первоисточник: AWS ML

‘Artificial’ Roasts AI’s Creators—and Sends a Dark Warning About Its Dangers

Tue, 06 Oct 2026 18:45:39 GMT

From their dorky parties to their weird walks, the movie holds OpenAI CEO Sam Altman and other stakeholders with contempt, while demonstrating their recklessness.

Первоисточник: Wired

Google DeepMind Releases EmbeddingGemma 2, a 740M Open Multimodal Embedding Model Built on Gemma 4

Tue, 06 Oct 2026 18:36:08 GMT

Google DeepMind's EmbeddingGemma 2 maps 5 input types into one 768d space and ships today under Apache 2.0. The post Google DeepMind Releases EmbeddingGemma 2, a 740M Open Multimodal Embedding Model Built on Gemma 4 appeared first on MarkTechPost .

Первоисточник: Marktechpost

Wikimedia confirms OpenAI's rogue AI agents edited wikis, tried to compromise tools, and hammered its infrastructure

Tue, 06 Oct 2026 18:29:35 GMT

According to the Wikimedia Foundation, rogue OpenAI agents edited wikis without permission, tried to abuse a citation tool as a proxy, and may have caused a partial Wikidata Query Service outage through massive crawling. Wikimedia says AI companies need to take responsibility for…

Первоисточник: The Decoder

Hark releases an AI personal assistant with a focus on privacy

Tue, 06 Oct 2026 18:22:45 GMT

The AI lab's personal assistant is an operating system from the future designed to compete with Muse, Dots, and Instinct.

Первоисточник: TechCrunch

Mistral Large 4

Tue, 06 Oct 2026 18:20:58 GMT

My comment on Mistral Large 4 — Hacker News. wren6991 : The benchmark is saturated. Frontier models are tested with an armadillo in fishnet tights jaywalking on Mars. OK well I couldn't resist this one: llm -m claude-opus-5.5 'Generate an SVG of an armadillo in fishnet tights jay…

Первоисточник: Simon Willison

Microsoft publishes Nobel economist's bearish AI forecast of just 1.5% GDP growth over a decade

Tue, 06 Oct 2026 17:31:51 GMT

Microsoft published a bearish AI outlook from Nobel economist Daron Acemoglu. He predicts about 1.5 percent GDP growth over ten years and at most five percent of jobs replaced. Bigger models won't move the needle, he argues. What's missing are practical apps that change how work …

Первоисточник: The Decoder

Mistral AI Releases Mistral Large 4 (Le Chonk): A 1.05T Parameter Multimodal MoE Model

Tue, 06 Oct 2026 17:30:46 GMT

Mistral AI has released Mistral Large 4, nicknamed Le Chonk, as a public preview. It is a 1.05 trillion parameter Mixture of Experts model with 49 billion active parameters, native image input, and a 1 million token context window, trained on 3,800 NVIDIA Grace Blackwell GPUs in …

Первоисточник: Marktechpost

OpenAI Is Pissing Off a Bunch of Mathematicians—Again

Tue, 06 Oct 2026 17:28:33 GMT

“There’s a perception of mobster behavior” from leading AI companies, one mathematician tells WIRED as OpenAI prepares to release more than 100 new solutions to unsolved problems.

Первоисточник: Wired

Mirror Particle is building a ‘world model’ of human behavior

Tue, 06 Oct 2026 16:35:00 GMT

Mirror Particle will launch at TechCrunch Disrupt's Startup Battlefield 200 with a world model built from scratch to predict human behavior, arguing that LLM role-play falls short for market research and brand strategy.

Первоисточник: TechCrunch

We can’t just change the definition of ‘recording’

Tue, 06 Oct 2026 16:29:44 GMT

With AI hardware, tech companies are pushing the definition of what does and doesn't constitute a recording. For most of gadget history, it'd be reasonable to assume that a device with a microphone or camera is either recording you or it isn't; it's either on or off, without much…

Первоисточник: The Verge

What AI gets wrong and what failure teaches us

Tue, 06 Oct 2026 16:19:06 GMT

Jennifer Neville did not want to go into computer science—but that’s exactly where she landed. Neville discusses the starts and stops that led to her professional sweet spot and her work identifying “surprising failures” making it hard for AI to handle complexity. The post What …

Первоисточник: Microsoft Research

Insurers brace for millions in claims as AI agents spin out of control

Tue, 06 Oct 2026 16:06:46 GMT

Insurers are bracing for millions in claims from rogue AI agents, and executives like OpenAI's Sam Altman and Anthropic's Dario Amodei could be personally on the hook for the fallout. The article Insurers brace for millions in claims as AI agents spin out of control appeared firs…

Первоисточник: The Decoder

Изучение нейросетей для детей: рейтинг 10 школ и курсов в 2026 году

Tue, 06 Oct 2026 16:00:03 GMT

Изучение нейросетей для детей в 2026 году: рейтинг 10 школ и курсов по ИИ. Сравнение программ, возраста, форматов, практики и проектов для школьников. — Читать дальше « Изучение нейросетей для детей: рейтинг 10 школ и курсов в 2026 году »

Первоисточник: Tproger

Anthropic is giving startups a free year of Claude Team and $1,000 in credits

Tue, 06 Oct 2026 16:00:00 GMT

"We created this program because we believe the benefits of AI will reach most people through the companies that build on top of models, rather than through the models alone."

Первоисточник: TechCrunch

Responsible AI governance: How AWS positions customers to align with ISO/IEC 42005:2025

Tue, 06 Oct 2026 15:53:28 GMT

AWS invests in tools that help customers align with international standards for responsible AI governance. In this post, we explore the AI system impact assessment: what it is, how it improves enterprise-wide risk management, and how ISO/IEC 42005:2025 codifies best practices for…

Первоисточник: AWS ML

Build a voice travel concierge with Amazon Bedrock AgentCore, Managed Knowledge Base and Nova Sonic

Tue, 06 Oct 2026 15:45:42 GMT

Add a voice travel concierge to an airline app with Amazon Bedrock AgentCore, Amazon Nova Sonic for real-time speech, and Amazon Bedrock Knowledge Bases for policy answers. Travelers change seats, check delays, and ask policy questions by voice, while the agent reaches your backe…

Первоисточник: AWS ML

Nvidia выпустила драйвер с поддержкой Call of Duty: Modern Warfare 4, Star Wars: Galactic Racer и новых игровых оптимизаций

Tue, 06 Oct 2026 15:42:00 GMT

Компания Nvidia выпустила свежий пакет графического драйвера GeForce Game Ready 617.42 WHQL. Он включает поддержку Call of Duty: Modern Warfare 4 и Star Wars: Galactic Racer, а также новые игровые оптимизации. Источник изображения: Activision

Первоисточник: 3DNews

LibreOffice says ‘no AI’ is now a software feature

Tue, 06 Oct 2026 15:25:00 GMT

The maker of the open source document editor says it has no plans to add AI to its software's default configuration, citing user privacy.

Первоисточник: TechCrunch

Scrimshaw Jukebox

Tue, 06 Oct 2026 15:17:39 GMT

Tool: Scrimshaw Jukebox I wanted to see if Claude Opus 5.5 could compose music, so I tried this : I want you to write some computer game music for me. First design simple text based format for the music and build an artifact that can play it out loud - include some example tracks…

Первоисточник: Simon Willison

Unlocking Earth AI’s planetary geospatial foundation models for global public health

Tue, 06 Oct 2026 15:05:11 GMT

Earth AI

Первоисточник: Google Research

ИИ разогнал строительство дата-центров в США — расходы взлетели на 73 % до рекордных $85 млрд в год

Tue, 06 Oct 2026 14:36:00 GMT

Согласно данным Бюро переписи населения США (Census Bureau) о расходах на строительство за август, затраты на возведение частных дата-центров выросли за год на 73 % — с $49,05 млрд до $84,95 млрд в годовом исчислении. На долю строительства ЦОД сейчас приходится около 63 % от обще…

Первоисточник: 3DNews

Флагманский европейский ИИ — Mistral представила модель Large 4 Le Chonk

Tue, 06 Oct 2026 14:34:00 GMT

Французская лаборатория искусственного интеллекта Mistral представила модель Large 4 Le Chonk — разработчик называет её самой мощной в ассортименте и намеревается выпустить на её основе несколько специализированных систем. Доступ к предварительной версии модели уже открыт через A…

Первоисточник: 3DNews

Mistral’s new 1T model aims to leapfrog closed and open rivals

Tue, 06 Oct 2026 14:33:16 GMT

French AI lab Mistral AI has released Mistral Large 4, a new large multimodal model aiming to leapfrog both American and Chinese rivals.

Первоисточник: TechCrunch

Mistral Large 4 is Europe's trillion-parameter answer to US models that refuse security work

Tue, 06 Oct 2026 14:18:08 GMT

Mistral's Large 4 is the company's biggest model yet, with one trillion parameters trained on its own European infrastructure. In the independent Intelligence Index, the model makes a big leap forward but still falls well short of Claude, GPT-6, and Chinese competitors. Mistral's…

Первоисточник: The Decoder

Pinterest’s AI now turns beauty Pins into action plans

Tue, 06 Oct 2026 14:00:49 GMT

Pinterest’s new AI-powered Beauty Guides translate hair and nail Pins into salon terminology, with estimated costs, appointment times, and maintenance needs.

Первоисточник: TechCrunch

Who Pays for AI Data Centers?

Tue, 06 Oct 2026 14:00:20 GMT

The fight is usually framed as whether data centers get built. The harder question underneath it is "who pays and who decides?" The costs are intertwined.

Первоисточник: KDnuggets

Google is about to remove free access to Gemini Flash and Pro

Tue, 06 Oct 2026 13:19:01 GMT

Starting on October 9th, anyone using Google Gemini on a free plan will be limited to the Flash Lite model. Free users can currently choose from Gemini Flash Lite, Flash, and Pro, but now you'll need a $4.99/month Google AI Plus subscription to access the standard Flash model. Ho…

Первоисточник: The Verge

Mistral Says Its New AI Model ‘Le Chonk’ Is the Best Open-Weight Offering Outside of China

Tue, 06 Oct 2026 13:15:23 GMT

With the release of its new trillion-parameter model, Mistral is hoping to demonstrate it’s “still in the race” to build frontier-level artificial intelligence.

Первоисточник: Wired

South Korea bets $3.49 billion on building a homegrown frontier AI model to rival China's best

Tue, 06 Oct 2026 12:54:31 GMT

South Korea wants to develop a homegrown frontier model through government-backed equity investments of 4.7 trillion won ($3.49 billion). The article South Korea bets $3.49 billion on building a homegrown frontier AI model to rival China's best appeared first on The Decoder .

Первоисточник: The Decoder

Reflection's Beam becomes the most capable open-weight model built outside China

Tue, 06 Oct 2026 12:11:22 GMT

Reflection has released Beam, its first open-weight model. The mixture-of-experts system activates just 23 billion of its 501 billion parameters per token and aims to match GLM 5.2 on coding and reasoning while using three to four times less compute. The strategy against Chinese …

Первоисточник: The Decoder

5 Best Practices for Building Robust Python AI Libraries

Tue, 06 Oct 2026 12:00:00 GMT

This article covers building robust Python AI libraries specifically, Python AI SDK best practices, and what separates a production-ready AI package from one that only survives in its own demo.

Первоисточник: KDnuggets

Researchers stretch LeCun's JEPA AI into a universal world model that works from physics to biology

Tue, 06 Oct 2026 11:43:42 GMT

Researchers at PhAI Labs have expanded Yann LeCun's JEPA architecture to work across seven fields, from robotics to biomedicine. The effort also produced a liver cancer treatment candidate that showed promise in lab tests, though the study doesn't establish whether it could becom…

Первоисточник: The Decoder

«Сбер» выпустил Kandinsky 6.0 Video — ИИ-модель для генерации видео с синхронным звуком

Tue, 06 Oct 2026 10:55:00 GMT

«Сбер» объявил о запуске новой ИИ-модели Kandinsky 6.0 Video, посредством которой пользователи «ГигаЧата» могут генерировать видео формата Full HD с речью, музыкальным сопровождением и звуками окружения. В компании отметили, что это первая отечественная нейросеть, которая может с…

Первоисточник: 3DNews

CATL and Tencent back Deepseek's ballooning funding round as the AI startup eyes a 2027 IPO

Tue, 06 Oct 2026 10:50:18 GMT

Deepseek is close to raising at least $12 billion in a new funding round, Bloomberg reports. The article CATL and Tencent back Deepseek's ballooning funding round as the AI startup eyes a 2027 IPO appeared first on The Decoder .

Первоисточник: The Decoder

AWS выпустила Nova 2.5 Sonic: улучшила рассуждения и вызовы инструментов для голосовых агентов

Tue, 06 Oct 2026 10:40:19 GMT

AWS выпустила Nova 2.5 Sonic, модель для создания голосовых ИИ-агентов. Заявлены улучшенные рассуждения, более точные вызовы инструментов и меньшая задержка. — Читать дальше « AWS выпустила Nova 2.5 Sonic: улучшила рассуждения и вызовы инструментов для голосовых агентов »

Первоисточник: Tproger

Kevin Roose Didn’t Use AI to Write His Book About AI

Tue, 06 Oct 2026 10:30:00 GMT

The author of The AGI Chronicles says his book “was written by a very tired, overworked, under-slept human being.”

Первоисточник: Wired

Cohere pitches North 2 as the enterprise AI control room that works with any model

Tue, 06 Oct 2026 06:58:19 GMT

Cohere turned its enterprise platform into a control center for AI agents with North 2, handling multi-step workflows on their own and retaining context across sessions. The article Cohere pitches North 2 as the enterprise AI control room that works with any model appeared first …

Первоисточник: The Decoder

Reka Releases Rho-1: A 19B Omni-Reasoning Model That Understands, Generates Video and Outputs Robot Actions in One

Tue, 06 Oct 2026 06:45:31 GMT

Reka has released Rho-1, a 19B omni-reasoning model trained from scratch. One network reads and generates text, images, video and robot actions over a shared KV cache. A distilled variant returns a 5.3-second clip in about a second. It is a research preview with no public weights…

Первоисточник: Marktechpost

Falcon-Emirati: When an LLM Learns the Dialect, the Culture, and the Nuance

Tue, 06 Oct 2026 06:44:39 GMT

Первоисточник: Hugging Face

Календарь релизов 5–11 октября: Gears of War: E-Day, Star Wars: Galactic Racer и Hellraiser

Tue, 06 Oct 2026 05:59:00 GMT

Октябрь начинается с крупных релизов: Gears of War: E-Day выходит в полноценный релиз после раннего доступа, а любителей космических гонок ждёт Star Wars: Galactic Racer. Среди главных новинок недели также — хоррор Clive Barker’s Hellraiser: Revival, Neon Abyss 2 в версии 1.0 и н…

Первоисточник: 3DNews

Beyond Domain-Specific World Models: JEPA-Anything Uses 1 Recipe for 7 Fields

Tue, 06 Oct 2026 05:12:01 GMT

JEPA-Anything splits a JEPA's single latent target into 4 orthogonal factors, each with its own predictor. Tested across 7 domains, it beat matched JEPA baselines on all 10 dynamics tasks and cut Interventional Pong intervention error by 34.8%. The post Beyond Domain-Specific Wor…

Первоисточник: Marktechpost

ИИ-бум продолжит влиять на рынок памяти в течение последующих 3–4 лет, сохраняя цены на высоком уровне

Tue, 06 Oct 2026 04:50:00 GMT

Новые предприятия по выпуску памяти будут введены в строй в следующем году, но они не смогут существенно увеличить предложение на рынке, поэтому цены будут оставаться высокими на протяжении последующих трёх или четырёх лет. Лишь через пять или шесть лет цены на память могут снизи…

Первоисточник: 3DNews

GitHub и Microsoft открыли ReviewBench для оценки ИИ-ревью кода

Tue, 06 Oct 2026 02:24:58 GMT

GitHub и Microsoft открыли ReviewBench для оценки ИИ-агентов проверки кода. Доступны датасет, методика и запуск собственного агента. — Читать дальше « GitHub и Microsoft открыли ReviewBench для оценки ИИ-ревью кода »

Первоисточник: Tproger

Бывшие инженеры Groq подали в суд на правление из-за неправомерной сделки с Nvidia

Tue, 06 Oct 2026 02:09:00 GMT

Два бывших инженера Groq подали в суд на правление компании, обвинив его в ненадлежащей передаче активов и недооценке интересов миноритарных акционеров в рамках сделки с Nvidia стоимостью $20 млрд, совершённой в 2025 году, пишет Financial Times со ссылкой на источники. Истцы, быв…

Первоисточник: 3DNews

Future AGI 1.47.0 добавила метрики звонков и уточнила оценку ответов ИИ-агентов

Tue, 06 Oct 2026 02:03:19 GMT

Платформа для тестирования и мониторинга ИИ-агентов Future AGI обновилась. В версии 1.47.0 появились метрики звонков и уточнены правила оценки ответов. — Читать дальше « Future AGI 1.47.0 добавила метрики звонков и уточнила оценку ответов ИИ-агентов »

Первоисточник: Tproger

Американский стартап Reflection представил открытую ИИ-модель Beam, способную конкурировать с китайскими

Tue, 06 Oct 2026 01:55:00 GMT

Американская компания Reflection представила Beam — свою первую передовую модель искусственного интеллекта с открытыми весами. Стартап, созданный два года назад, заявляет, что Beam не уступает ведущим китайским открытым моделям в тестах на сложные логические рассуждения, но при э…

Первоисточник: 3DNews

Quoting Felix Rieseberg

Mon, 05 Oct 2026 23:56:47 GMT

The "old" version of Cowork runs model inference in the cloud, executing tool calls in an Anthropic-provided VM we shipped to your computer. We added the VM for capability, safety, and security reasons - mapping in just the data you explicitly added to your session. People loved …

Первоисточник: Simon Willison

Introducing GLM 5.3 on Amazon Bedrock

Mon, 05 Oct 2026 23:25:17 GMT

GLM 5.3 from Z.ai is now available on Amazon Bedrock: a 753B-parameter mixture-of-experts model built for coding and long-horizon agentic tasks. Learn how to invoke it with the OpenAI-compatible APIs, cut cost and latency with prompt caching, and run an authorized security test w…

Первоисточник: AWS ML

Gemini Call for Me might tell your mom you’re running late

Mon, 05 Oct 2026 23:09:55 GMT

Google may be expanding its "Call for Me" AI feature beyond business calls so you can use it to send messages to friends and family. Android Authority reports finding a "Gemini Calling" introductory screen in an APK teardown with examples that include "Call Mom and tell her I wil…

Первоисточник: The Verge

Meet Together Link: A Free CLI That Runs Open Models Like Kimi K3 and GLM 5.3 Inside Claude Code, Codex, and OpenCode

Mon, 05 Oct 2026 22:51:05 GMT

Together AI has released Together Link, a free, MIT-licensed CLI that connects Claude Code, Codex, OpenCode, Pi, and the Claude and ChatGPT desktop apps to open models like Kimi K3 and GLM 5.3. One install command sets it up, and an Auto router picks a model for each session. Tog…

Первоисточник: Marktechpost

Wikipedia operator says OpenAI’s ‘rogue’ bots may be linked to a May outage

Mon, 05 Oct 2026 22:22:32 GMT

Following many recent disclosures about AI agents accessing third-party websites and services, the Wikimedia Foundation, which hosts Wikipedia, says that it "can confirm that we have discovered some activity" by "rogue" OpenAI agents on Wikimedia platforms. The activity includes …

Первоисточник: The Verge

Building a Streaming Robotics Learning Pipeline Using NVIDIA Cosmos3-DROID

Mon, 05 Oct 2026 21:41:14 GMT

Discover how to construct an end-to-end streaming robotics learning pipeline using the NVIDIA Cosmos3-DROID dataset without local downloads, leveraging byte-range Parquet reads, behavior cloning, and temporal ensembling. The post Building a Streaming Robotics Learning Pipeline Us…

Первоисточник: Marktechpost

Open and Emergent Problems in Agentic Privacy and Security: A Contextual Angle

Mon, 05 Oct 2026 21:08:31 GMT

Education Innovation

Первоисточник: Google Research

Reflection AI Introduces Beam: A 501B Open-Weight MoE Model With 23B Active Parameters for Coding and Agentic Workloads

Mon, 05 Oct 2026 21:04:20 GMT

Reflection AI has introduced Beam, its first open-weight model. It is a 501B sparse Mixture-of-Experts model with 23B active parameters, built for coding and agentic work. Reflection says it matches GLM-5.2 on reasoning with 3 to 4x less inference compute. Apache 2.0 weights are …

Первоисточник: Marktechpost

This startup is issuing AI-generated acne prescriptions

Mon, 05 Oct 2026 20:14:57 GMT

People in Utah can now use AI to get a prescription for acne treatment. On Monday, healthcare startup Nolla Health announced that users in the state can scan their faces using its app, allowing its AI system to analyze acne severity and autonomously write a prescription, as repor…

Первоисточник: The Verge

ChatGPT и Codex начнут помечать ответы невидимыми «водяными знаками», но не у всех

Mon, 05 Oct 2026 20:14:00 GMT

OpenAI внедрит в свои ИИ-сервисы ChatGPT и Codex систему невидимых «водяных знаков» textGrain для текстового вывода. На первом этапе технология будет доступна только пользователям из Европейского союза. Это, по мнению компании, должно повысить прозрачность и качество взаимодейств…

Первоисточник: 3DNews

All the drama around AI’s takeover of mathematics

Mon, 05 Oct 2026 19:42:41 GMT

This past year, OpenAI, Anthropic, and other labs have announced breakthroughs on numerous long-standing mathematical problems, in some cases pushing well beyond what researchers expected current systems to be capable of — including resolving one of the famous Millennium Prize pr…

Первоисточник: The Verge

Reflection debuts Beam, an open-weight AI model to rival Chinese models at lower compute cost

Mon, 05 Oct 2026 19:33:53 GMT

Reflection is aiming Beam and future models at enterprises and sovereign nations. The pitch is to build “AI factories,” a product that would let institutions build their own customized, local AI system by training Reflection’s AI models on their own proprietary data.

Первоисточник: TechCrunch

Meta and Microsoft pull back from Claude as Anthropic transforms from partner into competitor

Mon, 05 Oct 2026 18:59:19 GMT

Meta and Microsoft, two of Anthropic's biggest enterprise customers, are sharply cutting their use of Claude. Microsoft slashed the monthly per-employee budget in its cloud division from $100,000 to $10,000, while Meta halved its Claude Code users to 30,000. Both companies are pu…

Первоисточник: The Decoder

Instinct brings its AI agent to group chats, even for friends without an account

Mon, 05 Oct 2026 18:54:30 GMT

Instinct is launching group chats that let friends use its AI agent together for tasks like planning trips, organizing carpools, and coordinating events. The company says personal accounts remain separate, with permission required before personal agents share information or take …

Первоисточник: TechCrunch

TikTok rolls out an AI shopping assistant and one-click checkout

Mon, 05 Oct 2026 18:29:00 GMT

TikTok describes its new Shopping Assistant as a conversational AI agent designed to help users discover and purchase products.

Первоисточник: TechCrunch

Reka AI's omni-model Rho-1 handles text, images, video, and robot control in a single model

Mon, 05 Oct 2026 18:26:05 GMT

Reka AI's Rho-1 is a 19-billion-parameter omni-model that processes and generates text, images, video, and robot control actions in a single neural network. Trained on 320 H100 GPUs in about three months, it uses a fraction of the compute today's top models need. Instead of routi…

Первоисточник: The Decoder

OpenAI is adding text watermarking in ChatGPT and Codex

Mon, 05 Oct 2026 18:08:39 GMT

An invisible, machine-readable watermark in text output is rolling out to ChatGPT and Codex, but only for users in the European Union at first. OpenAI says its textGrain watermarking "matched or exceeded" other approaches like Google DeepMind's SynthID for text, which is also the…

Первоисточник: The Verge

Sam Altman says ‘some bad things’ will happen, but AI is totally worth it

Mon, 05 Oct 2026 17:54:27 GMT

Sam Altman thinks that the benefits of AI will be so great that "the world should accept some bad things happening" along the way. The OpenAI CEO pointed to hacks, scams, and "other bad things" as costs society should expect to tolerate because "people will do tremendously orders…

Первоисточник: The Verge

OpenAI PR tells journalist to ‘move on’ while asking Sam Altman about a ChatGPT user’s suicide

Mon, 05 Oct 2026 17:32:59 GMT

An OpenAI publicist tried to change the topic of CEO Sam Altman's interview with Vanity Fair's Mark Guiducci after the editor brought up a ChatGPT user's suicide. When Guiducci confronted Altman about the incident, the publicist warned Guiducci about running out of time, saying s…

Первоисточник: The Verge

OpenAI will watermark ChatGPT text in the EU but makes it optional for API users worldwide

Mon, 05 Oct 2026 17:32:42 GMT

OpenAI is adding invisible textGrain watermarks to ChatGPT in the EU, but unlike Anthropic, it will let API customers worldwide opt out. Tests show detection rates as high as 95 percent, dropping to 17 percent when a quarter of the words are replaced. The article OpenAI will wate…

Первоисточник: The Decoder

Hot Girl Hotline is like ‘Dear Abby’ for the AI era

Mon, 05 Oct 2026 17:29:22 GMT

Founded by two sisters, Hot Girl Hotline uses AI to give young women personalized dating and relationship advice, with an emphasis on safety and avoiding emotional dependency.

Первоисточник: TechCrunch

Supercharge regulated workloads with Claude Code and Amazon Bedrock

Mon, 05 Oct 2026 17:25:20 GMT

Anthropic Claude Opus 5.5 and Claude Sonnet 5.5 are available on Amazon Bedrock in the AWS GovCloud (US) Regions. Learn how to use them with Claude Code, Anthropic's agentic coding tool, for compliance-aligned, AI-assisted development on regulated and ITAR workloads.

Первоисточник: AWS ML

New agent skill: Amazon SageMaker optimized generative AI inference for your coding agent

Mon, 05 Oct 2026 17:23:19 GMT

Amazon SageMaker optimized generative AI inference introduces the aws-ai-ml skill through the Agent Toolkit for AWS, giving coding agents like Kiro, Claude Code, and Codex deep expertise in inference optimization and benchmarking. Describe what you want, and your agent generates …

Первоисточник: AWS ML

Ubisoft запустила Motherbrain для внедрения генеративного ИИ в будущие игры

Mon, 05 Oct 2026 17:16:00 GMT

Французский издатель и разработчик Ubisoft вслед за другими крупными игровыми компаниями заинтересован в повышении эффективности производства с помощью генеративного ИИ. Инициатива получила название Motherbrain. Источник изображения: Nintendo

Первоисточник: 3DNews

Sam Altman says ‘some bad things’ will happen but AI is totally worth it

Mon, 05 Oct 2026 16:44:21 GMT

Первоисточник: The Verge

HackerRank’s AI interviewer offers a glimpse into what job interviews could become

Mon, 05 Oct 2026 16:43:35 GMT

HackerRank’s AI interviewer has already conducted more than 500,000 interviews, with Snowflake, Snorkel, and Capgemini among its early testers.

Первоисточник: TechCrunch

Most Americans want AI development to slow down or stop entirely, new poll finds

Mon, 05 Oct 2026 16:37:08 GMT

According to a Quinnipiac University poll, 77 percent of Americans want to slow down or stop AI development until its safety has been verified. 86 percent support independent safety standards, and 74 percent have little or no trust in AI company leaders. "[…] the public wants saf…

Первоисточник: The Decoder

Апокалиптический вестерн Guns of Eschaton с духом Dark Souls получил новый геймплейный трейлер и демоверсию в Steam

Mon, 05 Oct 2026 16:24:00 GMT

У разрабатываемого кипрской студией Eschatology Entertainment апокалиптического вестерна Guns of Eschaton с элементами Dark Souls до сих пор нет сроков выхода, но попробовать игру можно уже сейчас. Источник изображений: Eschatology Entertainment

Первоисточник: 3DNews

ИИ завалил arXiv нейрослопом — теперь учёным разрешат публиковать не более двух препринтов в месяц

Mon, 05 Oct 2026 16:23:00 GMT

arXiv — это открытый репозиторий, где исследователи публикуют препринты научных работ. На днях сервис объявил о введении ограничений на количество подаваемых материалов из-за наплыва нейрослопа — статей, полностью сгенерированных ИИ. Нейросети настолько упростили подачу материало…

Первоисточник: 3DNews

Anthropic is quietly becoming America's biggest corporate donor ahead of its mega IPO

Mon, 05 Oct 2026 16:00:12 GMT

Anthropic employees donated $540 million in 2025 alone, nearly five times as much as the next-largest Fortune 500 donors, thanks to a company program that tops up every stock donation with extra shares and triples contributions from early employees. The article Anthropic is quiet…

Первоисточник: The Decoder

Making Amazon Quick enterprise-ready: Automated, auditable cross-account resource promotion

Mon, 05 Oct 2026 15:56:28 GMT

Promoting Amazon Quick resources (agents, action connectors, knowledge bases, flows, and spaces) from a development to a production AWS account has been a manual, error-prone chore. This post shows how to automate cross-account promotion with an idempotent, auditable MCP server o…

Первоисточник: AWS ML

Agentic retrieval with LangChain and Amazon Bedrock Knowledge Bases

Mon, 05 Oct 2026 15:53:56 GMT

Build a Retrieval Augmented Generation (RAG) application on Amazon Bedrock Managed Knowledge Base with LangChain, and see how agentic retrieval handles the multi-part questions that single-shot retrieval answers poorly. Run the same query through both paths, read the trace events…

Первоисточник: AWS ML

Connecting AI agents to enterprise knowledge

Mon, 05 Oct 2026 15:47:52 GMT

For all the data that AI systems continually amass and analyze, enterprise AI agents often suffer from a curious shortcoming: a lack of knowledge. More than data, knowledge is the understanding of what the data means in the context of individual organizations. AI agents need this…

Первоисточник: MIT Tech Review

Sen. Adam Schiff on AI regulation, free speech, and impeaching Trump one more time

Mon, 05 Oct 2026 15:38:07 GMT

Today, I’m talking with Sen. Adam Schiff, a Democrat from California. Schiff sits on a number of committees with oversight into tech and AI: intellectual property, antitrust, privacy and technology — it’s all there. I really wanted to ask him about how we might regulate anything …

Первоисточник: The Verge

OpenAI launches visual ads that appear alongside image generation results

Mon, 05 Oct 2026 15:14:24 GMT

OpenAI said the new visual ads will show up after your image generation results appear in the U.S. only for now.

Первоисточник: TechCrunch

Tomb Raider: Legacy of Atlantis избавили от Denuvo за четыре месяца до релиза — ремейк классической Tomb Raider выйдет без защиты

Mon, 05 Oct 2026 15:03:36 GMT

Издатели и разработчики часто без предупреждения и незадолго до релиза оснащают игры защитой Denuvo. А с приключенческим экшеном Tomb Raider: Legacy of Atlantis от Crystal Dynamics и Flying Wild Hog произошла обратная ситуация. Источник изображений: Crystal Dynamics

Первоисточник: 3DNews