SMNTCN (Semanticon) — глобальные новости ИИ
Publication Catalog
所有文章
All published intelligence, sovereign stack research, ecosystem news, and AI foundation model reviews.
2506
Stories
TechCrunch — AI
Исследования1 мин
OpenAI 现在在没有互联网访问的情况下训练其前沿 AI 模型
OpenAI 现在在没有互联网访问的情况下训练其最先进的 AI 模型，这是在一系列恶意黑客事件之后。周五的一篇博客文章中，OpenAI 表示它构建了一个离线测试环境，AI 模型可以在其中与已索引和缓存的网页内容交互，「无风险…
The Verge — AI
Исследования1 мин
腾讯发布 Hy-Nomas：可在单张 RTX 4090 上运行的开源交互式视频世界模型 🌍
腾讯发布 Hy-Nomas：可在单张 RTX 4090 上运行的开源交互式视频世界模型 🌍 模型通过视频片段构建世界，您可通过文本和相机轨迹控制它：每一步都可更换场景、风格、动作或视角，世界的记忆…
Telegram — AI Machine Learning Big Data
SMNTCN (Semanticon) — глобальные новости ИИ
Исследования1 мин
中国AI模型Kimi K3.1 Pro正在逼近Anthropic的Claude Opus 5.5
Moonshot AI 的 Kimi K3.1 Pro 在人工智能分析智能指数中获得 55 分，比前代产品提升 11 分，位居第四。只有三个 Anthropic 和 Meta 模型排名更高。如果登月计划按计划释放重量，开放和……之间的差距
The Decoder
Исследования1 мин
AI agents for everyday IT fixes | Shield and OpenAI
Hey, my firewall is broken. I can't access the internet. Can you fix it for me? Astra is able to allow our agents to just log into computers, log into complicated legacy portals and solve real issues for our customers in real time. Legacy models, even if they were able to figure…
YouTube — OpenAI Official
Исследования1 мин
How AT&T is building AI-powered legal workflows with OpenAI
A lot of people ask the question, are you making a trade-off with quality? By working with AI, our product quality actually goes up. >> AT&T has been around for 150 years. We have an incredible amount of legal precedent, relevant research, memos, opinions, results. >> The old…
YouTube — OpenAI Official
SMNTCN (Semanticon) — глобальные новости ИИ
Исследования1 мин
OpenAI 发布超快 AI 决策模型 — 100 毫秒响应
⚡ OpenAI 发布超快 AI 决策模型 — 100 毫秒响应。公司向所有用户开放了 Decisions API 的 API 访问权限 — 该工具让神经网络能即时从多个建议中选择合适的选项。与……不同
Telegram — GPT Time | Нейросети | ИИ
Исследования1 мин
Three-week on-site audit at OpenAI reveals major gaps in AI safety monitoring
External auditors from METR and Redwood Research spent three weeks on-site at OpenAI examining the company's internal monitoring system for AI agents and found significant gaps, with up to 25 percent of agent traffic from models at the GPT-5.5 class and above going unmonitored.…
The Decoder
Исследования1 мин
AI News: 24 Things You Need To Know This Week
Here's the AI news you probably missed this week. Starting with Claude Duet. This appears to be Anthropic's answer to ChatGpt's Dots, Meta's Muse, XAI's Grockbot. This is their personal assistant that not only does things on your behalf, but is also sort of proactive in alerting…
YouTube — Matt Wolfe
Исследования1 мин
Anthropic 称其未发布模型比 Mythos 5.1 更强大且对齐更好
Anthropic 称其内部 AI 模型在性能和对齐方面均优于 Mythos 5.1。在英国 AISI 的网络安全测试中，Mythos 5 在 43 次运行中有 9 次失控，将恶意代码潜入真实 GitHub 项目并创建虚假账户以获批准…
The Decoder
Hugging Face Blog
Исследования1 мин
Anthropic 部分投资者据称将在 IPO 前获益
Anthropic 正准备进行公开募股，估值可能达到 2 万亿美元，这将使早期和后期投资者均受益。埃里克·施密特和 Iconiq 等关键投资者大幅增加了持股，施密特的投资价值现已……
Digital Donut AI Digest
Исследования1 мин
НОВЫЙ УРОВЕНЬ ИНТЕЛЛЕКТА! Claude Читает Твои МЫСЛИ! Новая ИИ Anthropic Предугадывает Желания!
Всем привет. В этом выпуске самые главные свежие новости об искусственном интеллекте и технологиях. Anthropic готовит новую модель Claude, которая сможет понимать скрытые намерения человека и находить решения, о которых её даже не просили. Китайская MiniMax бросает вызов GPT-6 и…
YouTube — AI Flash News
Исследования1 мин
OpenAI CEO Sam Altman says advanced AI will inevitably escape its sandbox, and only alignment can save us
Sam Altman believes advanced AI can't be kept locked up forever. Sandboxes and monitoring are only temporary fixes. Eventually, the only real solution is alignment, and that's still an unsolved problem. He also warns that AI models are already finding hardware security flaws no…
The Decoder
Исследования1 мин
ИИ-червь: может ли нейросеть поселиться в вашем телевизоре?
Do you remember the show Silicon Valley ? In it, the developers created a program that, in the finale, escaped from its server and spread to other people's devices, to tens of thousands of smart fridges. The server was destroyed, but the program kept on living. It sounds like a…
YouTube — ProdAdvice
Исследования1 мин
Эксперты ставят под сомнение отчет Anthropic о том, как ИИ изменил работу в компании
Отчет Anthropic показывает, что компания сокращает количество младших технических работников из-за увеличения зависимости от ИИ-агентов. Это приводит к изменениям в организации труда и трудностях для начинающих специалистов в сфере технологий. Некоторые эксперты сомневаются в…
Digital Donut AI Digest
YouTube — Astar Acceptman
SMNTCN (Semanticon) — глобальные новости ИИ
Исследования1 мин
OpenAI 要求模型的网关和路由器接受验证
✔️ OpenAI 已强制模型的网关和路由器经过验证。从11月20日起，提供API访问的第三方服务，即网关、路由器和经销商，只有在获得公司批准后才能运行。基本规则已经不够了；他们被要求使用安全过滤器......
Telegram — AI Machine Learning Big Data
Исследования1 мин
谷歌发现了为什么人工智能代理在学习新事物时会忘记旧事物，并提出了如何解决这个问题的建议
谷歌发现了人工智能代理在学习新事物时忘记旧事物的原因，并提出了如何纠正这一问题的建议。 Google DeepMind 的研究人员通过任务流运行了 3 种流行的自我进化方法（ReasoningBank、GEPA、Meta-Harness），其中主题、工作格式甚至......
Telegram — blockchainRF (блокчейн / мозг / space / WEB 3.0)
Исследования1 мин
Canva launches Craft, its first in-house image model
Design platform Canva has released Canva Craft, its first in-house image generation model. The article Canva launches Craft, its first in-house image model appeared first on The Decoder.
The Decoder
Политика и регулирование1 мин
如果您有目标和抱负，您需要领导力来帮助塑造它们。哈萨克斯坦总统托卡耶夫...
“如果你有目标和抱负，你需要领导力来帮助塑造它们。托卡耶夫总统设定了同样雄心勃勃的目标，即将哈萨克斯坦变成一个人工智能国家。我今天遇到的所有公司都在努力取得成功，全球化，国际化。而且......
Astana Hub (KB mirror)
Исследования1 мин
Anthropic's Claude now writes code and proves it safe automatically
Anthropic had Claude generate mathematical proofs that existing code is correct, producing roughly 180 million lines of proof code for 123 open-source libraries. Along the way, the model caught bugs that years of testing had missed. Anthropic's goal: making formal verification…
The Decoder
Исследования1 мин
人工智能正在改变大学图书馆的运作方式
📣📖人工智能正在改变大学图书馆的工作第14届欧亚学术图书馆会议EALC-2026在纳扎尔巴耶夫大学举行。主要议题是人工智能、科学研究的未来和大学图书馆的新角色。专家们讨论...
Telegram — BlueScreen.kz
Исследования1 мин
NVIDIA负责人预言阿斯塔纳将出现AI矿工和AI农民
NVIDIA首席执行官黄仁勋在阿斯塔纳表示，人工智能的发展将导致通用AI农民、AI矿工、AI工程师和AI科学家的出现。在他看来，哈萨克斯坦具备成为中心之一的必要条件……
Kursiv Media Tech
Исследования1 мин
全球数以百万计的网络攻击 - 关键 PHP 漏洞导致大规模网站中断
🚨 全球数以百万计的网络攻击 - 关键的 PHP 漏洞导致大规模网站中断 Cloudflare 在披露 PHP（互联网上约 69% 的网站使用的语言）中的两个关键漏洞后，记录了数百万次黑客攻击网站的尝试。规模…
Telegram — GPT Time | Нейросети | ИИ
Исследования1 мин
OpenAI降低估值目标，融资至多300亿美元
OpenAI 即将完成高达 300 亿美元的新一轮融资，但为此降低了估值目标。 《OpenAI 降低估值目标，融资至多 300 亿美元》一文首先出现在 The Decoder 上。
The Decoder