Перейти к содержимому

Утренняя сводка AI — 9 июля 2026

9 июля 2026 · 10 материалов из 5 фидов · проверено 262 статей за последние 36 часов


Источник: AI Insider

How AI Token Supply Shapes Prices and Investor Returns The economics of artificial intelligence are increasingly being written in a currency most investors have never priced before: the token. As enterprises push generative AI and autonomous agents into production, the volume of tokens consumed, and the supply available to meet that demand, has become one of the clearest signals of where value, cost, and eventually returns will land across the AI stack.


2. Microsoft Deploys In-House MAI Models to Cut AI Costs Amid Industry-Wide Spending Pullback

Заголовок раздела «2. Microsoft Deploys In-House MAI Models to Cut AI Costs Amid Industry-Wide Spending Pullback»

Источник: AI Insider

Microsoft Deploys In-House MAI Models to Cut AI Costs Amid Industry-Wide Spending Pullback Microsoft has reportedly begun relying more heavily on its own in-house MAI models rather than third-party AI from OpenAI and Anthropic, according to a Bloomberg report. The company has started using MAI models to handle a portion of user prompts within Excel and Word, marking a shift from its previous approach of powering large parts of Office 365 through OpenAI and Anthropic systems.


3. Anthropic Expands Claude Cowork to Mobile as Debate Grows Over Frontier Versus Open-Source AI Economics

Заголовок раздела «3. Anthropic Expands Claude Cowork to Mobile as Debate Grows Over Frontier Versus Open-Source AI Economics»

Источник: AI Insider

Anthropic this week extended Claude Cowork, its agentic tool for general knowledge work, to web and mobile platforms for Max subscribers, after launching as a desktop-only app in January. The update lets users start tasks from a desktop and monitor or retrieve results from their phone, reflecting Anthropic’s ambition to position Cowork as a background administrative agent rather than a coding tool.


4. IBM and Red Hat Expand Lightwell with New Offerings to Build the Trust Infrastructure for AI-Era Open Source

Заголовок раздела «4. IBM and Red Hat Expand Lightwell with New Offerings to Build the Trust Infrastructure for AI-Era Open Source»

Источник: AIwire

RALEIGH, N.C. and ARMONK, N.Y., July 8, 2026 — IBM and Red Hat today announced the commercial launch of Lightwell, delivering automated vulnerability remediation at scale through two offerings: Lightwell Network and Lightwell Clearinghouse Premier.


Источник: AIwire

by Mo Sarwat | July 8, 2026 Earlier this year, a wave of new AI coding capabilities crystallized a shift that had been building for months: LLMs began to look capable of replacing – not just augmenting – enterprise software. The idea that AI could “eat” SaaS moved from speculation to something closer to inevitability, sparking what many dubbed the “SaaSpocalypse.” However, we’re not there yet.


6. Frontier AI Model Power But 60X Cheaper? It’s Possible, Says Together AI

Заголовок раздела «6. Frontier AI Model Power But 60X Cheaper? It’s Possible, Says Together AI»

Источник: AIwire

Frontier models like ChatGPT, Gemini, and Claude are very powerful, but that capability comes at a steep cost. What if you could get the same level of artificial intelligence and reasoning from an AI model, but without shelling out millions in tokens?


Источник: Hugging Face Blog

Back to Articles TL;DR : The transformers vLLM backend is now as fast (or faster) than custom vLLM implementations for many LLM architectures. Model authors can automatically leverage their transformers implementations to get ultra fast vLLM inference, for free.


Источник: Hugging Face Blog

Back to Articles Why agentic AI needs open data, and why synthetic data is how we scale it. Image: Nemotron Post-Training v3 Prompt Atlas More Than Model Weights Building AI agents is hard, because the real world does not behave like a benchmark.


Источник: The Verge AI

OpenAI is overhauling ChatGPT’s voice mode with a new model that it says is more like “talking to another person.” The new GPT-Live-1 is designed to interrupt you less and will also wait for you to continue speaking if you pause mid-conversation. During a press briefing, OpenAI research lead Kundan Kumar called GPT-Live-1 the company’s “smartest voice model” yet.


10. Google updates Android Bench with new LLMs, but Gemini still lags behind

Заголовок раздела «10. Google updates Android Bench with new LLMs, but Gemini still lags behind»

Источник: Ars Technica

Code generation is emerging as one of the most popular applications for large language models (LLMs), but not all agents are equally good at all development tasks. Google created a benchmark earlier this year to evaluate how LLMs perform in Android app development, and Android Bench is getting a big update today.


Автосборка FreshRSS → Starlight · 09.07.2026 09:00