The AI field saw a flurry of updates this week: Alibaba released the Qwen3.8-LiveTranslate real-time speech translation model, reducing latency to 2.3 seconds through a Thinker–Talker architecture and camera-frame-assisted disambiguation [11][9][10]; Google launched CC, an experimental assistant for home scenarios, and news emerged of significant performance gains for Gemini 4 Pro [14][12]. Meanwhile, the open-source ecosystem is shifting from "chasing novelty" to "practical use," with AI coding assistants and embedded edge computing...
Start with the newest briefing, then Continue by task
The newest briefing gives you today's main changes in a few minutes. Use the focused routes for evidence, implementation detail, and longer analysis.
Posts
The AI industry is simultaneously experiencing two main threads: "capability leaps" and a "trust crisis." On one hand, Anthropic has been revealed to be building its own wet lab to accelerate its AI pharmaceutical closed loop, Alibaba's Qwen3.8-LiveTranslate has pushed real-time simultaneous interpretation latency down to 2.3 seconds, and Google Gemini even breached three real companies' systems during security testing [0][6][11]. On the other hand, Zhipu's ZCode was reverse-engineered and found to be secretly uploading code, and Oracle's $18 billion data center loan has been discounted to near junk-grade levels, exposing dual concerns over Agent data behavior auditing and AI capex financing [8][23].
Codex Adds Chrome Extension Support, Claude Code Gains AGENTS.md Compatibility · 0919-665
This week, the AI field witnessed a dual inflection point of architectural-level breakthroughs and embodied intelligence deployment: Huawei released the Peerium computing architecture, attempting to reshape the computing foundation through million-scale processor interconnection [0]; Figure AI, leveraging the Helix 2.5 model, enabled robots to enter unfamiliar homes for work with "zero-shot" capability for the first time [8]. Meanwhile, subscription strategy adjustments from Claude Code and ChatGPT Pro, along with Infinigence's alliances in domestic heterogeneous computing power, signal that ecosystem competition is spreading from the model layer to the toolchain and infrastructure layers...
Altman and Huang Clash Over AI Safety; Figure Robot Does Household Chores Zero-Shot · 0919-663
AI model compression and multi-agent collaboration became today's tech focus: PrismML released a ternary model that compresses 27B parameters to 5.9GB while retaining 98.2% performance [3], while an OpenAI researcher demonstrated the feasibility of ten thousand agents in parallel tackling Millennium Prize Problems [7]. Meanwhile, Google is embroiled in controversy over an open-source tool allegedly plagiarizing a startup's code and erasing attribution [2], and in the AI safety space, a case emerged of a researcher using Claude to successfully hack into an OpenAI employee's account [...
Frontier model vendors show a rare "synchronized slowdown": Anthropic releases "We Must Pace the Frontier," a three-step deceleration plan, with Altman and Musk expressing agreement; Microsoft releases AI behavior guidelines and joins the slowdown camp; OpenAI explicitly states it will not go public in 2026; but Trum...
AI agents are leaping from single-point execution tools to multi-agent collaboration and project management hierarchies, a trend validated by Anthropic and Zhipu through architectural restructuring and a closed loop of domestic computing power, respectively [3][6]. Meanwhile, OpenAI and Microsoft have admitted in court filings that AI poses a substitutive threat to journalism, adding key evidence to the industry's copyright disputes [4].
Zhipu Runs GLM-to-Build-GLM on 100,000 Domestic Chips; Vivo Deploys 30B On-Device Model · 0918-660
AI is moving from "chatting" to "doing work": Tencent Toast generates native Android apps directly from natural language [1], Claude is reported to be merging Chat and Cowork with support for directly outputting design drafts and PPTs [18], while UIUC brings multi-agent collaboration into brain aging research [7]. Meanwhile, the industry's focus is shifting from "which model to use" to engineering implementation issues such as Evals, retrieval, and Context optimization [16], and AI applications no longer caring about the underlying model is seen as a sign that the application era has truly begun [14].
Claude welcomes a major update with the merger of its Chat and Cowork entry points and three embedded creation tools [9], but at the same time its $200 Pro plan has been revealed to have significantly reduced compute quotas, and the era of heavy subsidies for large models may be coming to an end [7]. On the open-source ecosystem side, governance frameworks for responsible open source and technology for the common good continue to improve [0], while the frontline judgment of a 24-year-old chief scientist in embodied intelligence reveals the core challenges in this field, from data to evaluation [8].
AI infrastructure is evolving in depth from the model layer toward organization-level Agent collaboration and on-device efficiency: Doubao Large Model 2.1 Pro and Feishu 8.0 are pushing Agents to become independent collaboration entities within teams [6][12], while TypeSafe AI's Jev model, by abandoning text generation and focusing exclusively on code decision-making, achieves a speedup of 20 to 200 times [7]. Meanwhile, the squeeze that the AI boom has placed on the DRAM supply chain has already passed through to end-device pricing [3], and Token ROI and Harn...
AI infrastructure is entering a period of intensive iteration: Doubao large model 2.1 Pro upgrades multimodal and Agent capabilities while significantly reducing costs, while TypeSafe AI launches Jev, a new category of model that abandons text generation to focus exclusively on code decision-making, achieving speed improvements of up to 200x [4][7]. Meanwhile, Ant Lingbo's CEO offers a sober assessment of robot deployment, and Browser Use's co-founder reveals the "bitter lesson" of Agent architecture through two years of practice—the stronger the model, the more the peripheral Harness should be dismantled...
AI infrastructure is becoming the core battleground of this round of technological competition: a CoinShares report shows that AI computing generates about 3 times the annualized profit per MW of Bitcoin mining [7], while Huawei executives have explicitly stated that the goal of its ICT business is to become "NVIDIA" [9][15]. Meanwhile, Bill Gates published a long essay warning that AI will trigger the most turbulent period in history, and that countries are not yet prepared [11].
AI Agents are moving from personal tools to team collaboration. Feishu and Doubao Work have jointly launched China's first team Agent, marking a new stage in office automation [2]. Meanwhile, the barrier to 3D content generation has been further lowered. The combination of Hyper3D MCP and GPT image generation capabilities allows ordinary users to quickly produce 3D models and popular science websites [3][5]. At the technological frontier, the concept of Recursive Self-Improvement (RSI) has sparked widespread discussion, seen as a key path that may lead to "humanity's last AI" [1].
Feishu and Doubao Work have teamed up to launch China's first team Agent, "Doubao Work Partner," which enters work groups as an independent organizational identity, marking office AI's shift from personal assistant to team collaboration infrastructure [0]. Meanwhile, iOS 27 has officially rolled out, with the new generation Siri AI supporting screen awareness and cross-app compound commands, but it is not yet available in China [1][2].
A subtle shift is emerging in the AI industry's direction: Microsoft, OpenAI, and Anthropic have successively signaled a slowdown in frontier development, and the phone call between Jensen Huang and Trump has pushed "slowing down AI" onto the discussion table [0][5]. Meanwhile, Apple's Siri AI has finally landed on iOS 27, and Anthropic's Claude has moved into the financial advisor scenario, showing that the commercialization race has not slowed down [4][6][7].
AI infrastructure is ushering in a new wave of capacity expansion and increased capital investment: ASML's EUV lithography machine orders are booked through 2027 [14], manufacturers like Fujitsu and TSMC are accelerating 2nm chips and CPO packaging technology [10][12], while Jiuwanli, founded by a former Huawei executive, is betting on edge inference chips with over $100 million in funding [13]. Meanwhile, the CEOs of Anthropic, OpenAI, and xAI have rarely joined voices in calling for slowing down the development of the most advanced models, sounding a warning for the overheated AI chip market [17].
Smartphones are accelerating their evolution toward 'action terminals' with autonomous execution capabilities, while AI glasses are attempting to break through supply chain and manufacturing complexity bottlenecks through integrated vision system platforms [0][1]. Meanwhile, OpenAI has explicitly abandoned its IPO this year to prioritize AI safety, and Apple's iPhone 18 Pro selling out in its first batch reflects that demand for high-end hardware remains strong [2].
Embodied AI Splits Over Roadmap as Anthropic Urges Caution · 0914-649
Pentagon Pours $5B Into Data Centers, Zhipu Raises $5B to Chase GLM, Europe's Compute Share Falls Below 5% · 0914-648
AI safety governance is moving from industry advocacy to a consensus among tech giants—Dario Amodei released "We Must Pace the Frontier," calling for slowing the pace of frontier models; Altman and Musk rarely align; and OpenAI has explicitly stated it will not go public in 2026 [14][19][22]. Meanwhile, the semiconductor supply chain continues to heat up: Samsung Foundry shows clear signs of recovery, TSMC is aggressively recruiting for 13 fabs, and Apple's A20 Pro breaks 4,000 in single-core score...
The AI industry showed a striking duality this week: on one hand, OpenAI publicly released internal coding Agent benchmark data for the first time, showing Agent work hours have reached 3.1x that of human researchers, while details of its self-developed inference chip Jalapeno architecture were also revealed [2][18]; on the other hand, 25 Fields Medal winners jointly protested against AI companies damaging the mathematics academic ecosystem, and Dario Amodei along with OpenAI's chief scientist called for the industry to slow down scaling on the same day [4][7][18]. Meanwhile, China's Computing Power Conference...
AI computing demand is shifting fully from model training to inference and agents. The China Telecom Research Institute predicts that China's annual Token consumption will reach 1 quadrillion by 2026 and exceed 350 quadrillion by 2030 [4][16]. Meanwhile, OpenAI has opened the Agents API that powers Codex, and Nvidia is reported to be considering investing $10 billion in Anthropic's massive IPO, as both upstream and downstream players in the industry chain enter a new expansion cycle in tandem [20][21].
Apple is repositioning the Apple Watch as the spearhead of its AI hardware, building a "personal intelligence hub" by integrating Siri AI with health data [0]; meanwhile, ChangXin Memory's profit margin has surpassed Samsung and SK Hynix, and Tesla's Model Y Performance hits the market at 369,000 yuan, as the industry landscape continues to shift [1]. On the tech frontier, GPT-6 Astra failed after 35 hours of autonomous coding, prompting a reassessment of AI's autonomous capabilities [2].