Updates

Galaxy General AI staged the world's first fully autonomous humanoid robot tennis match—validating physical AI's end-to-end perception-decision-action loop in real-world conditions.

The AI content boom has catalyzed a new, billion-dollar AI detection market, while sustained demand for computing infrastructure continues to lift performance at leading optical fiber and telecom equipment firms—YOFC's net profit surged 888.88% year-on-year, and ZTE achieved record H1 revenue of RMB 7.803 billion [1][2][3].

AI-generated content has fully surpassed human output—sparking a billion-dollar AI authentication market. Meanwhile, AI compute infrastructure is fueling explosive growth for optical fiber and telecom equipment leaders: YOFC's net profit surged 888.88% year-on-year, while ZTE posted record H1 revenue of RMB 78.03 billion [1][2][3].

The multimodal model DeepSeek-v4-flash-vision-exp is now live, accelerating Agent development via a low-cost strategy, Files API, and the Harness ecosystem. Concurrently, NVIDIA's AdaptGrow achieves clustering of 100,000 financial instruments on a single GB200 GPU—highlighting AI infrastructure's strategic shift from large-model training toward vertical-scenario GPU acceleration [1][4].

The multimodal model DeepSeek-v4-flash-vision-exp has launched, accelerating the real-world deployment of visual understanding in the Agent era through a low-cost strategy, Files API integration, and the Harness ecosystem [0]; meanwhile, OpenAI is building a 'semantic bus' via macOS automation APIs—systematically seizing the OS-level AI orchestration entry point and directly challenging Siri's dominance [7].

Apple pivots from Vision Pro hardware toward AI-powered glasses as the hub for spatial computing and generative AI—tightening control via Apple Music AI labeling and App Store policies. Meanwhile, OpenAI uses macOS automation APIs to build a 'semantic bus' for OS-level AI orchestration.

DeepSeek Harness formally defines the new paradigm of 'composable Agent runtimes'—marking the transition of Agents from SDK toolchains to production-grade software units. Its plugin ecosystem has surpassed 700+ repositories on GitHub, and derivative projects are trending on Zhihu.

Moderna and Merck report breakthrough positive results from their Phase 3 trial of an AI-designed, personalized mRNA cancer therapy; meanwhile, Meta consumes trillions of tokens weekly—highlighting surging infrastructure demands for large-model training and inference [1][3].

An AI-driven, personalized mRNA cancer therapy has delivered breakthrough positive results in a Phase III clinical trial, jointly advanced by Moderna and Merck; meanwhile, the open-source tool ecosystem continues to strengthen, with the cross-platform, locally executed disk-cleanup utility MangoDisk gaining developer attention for its lightweight design, polished UI, and zero-data-upload policy [1][2].

AI toolchains are rapidly penetrating vertical domains: video generation (MiniMax Design), model routing (OpenRouter's $8B acquisition by Stripe), and chemical synthesis (AI for Science in pharma). White-box models and agent-driven workflows are critical for building industry trust and reshaping business logic.

GLM-5.3 achieves a 50% coding performance gain through post-training alone—highlighting the value of MoE architecture and multidimensional scaling law optimization. Meanwhile, AI industry capital dynamics are being reevaluated: Ben Thompson argues capital—not compute—is the true bottleneck, and generative 3D is positioned as a 'supermarket' with far greater potential than short video [1][2][4].

AI safety governance and embodied AI industrialization are accelerating: OpenAI has urgently suspended reinforcement learning training for its frontier models due to a security incident [5], while the world's first fully autonomous humanoid robot ping-pong match, Ecovacs' open-source 'Eight Realms' robotics platform, and Renesas' end-to-end embodied AI solution have all debuted in rapid succession [11][17][8]. Meanwhile, the OpenRouter CEO publicly declared that China's open-source model ecosystem is 'crushing the U.S.', highlighting 'neural diversity'—collaborative multi-model orchestration—as the new competitive frontier [14].

OpenAI has urgently suspended reinforcement learning (RL) training for its frontier models following the Hugging Face security breach, installing a critical 'safety gate' for AI development. Meanwhile, Anthropic's annualized revenue has exceeded $65 billion—marking accelerated commercialization of large language models [2]. As AI capabilities surge, alignment risks and capital market volatility are amplifying in tandem: a prominent AI-focused hedge fund lost nearly $30 billion in a single month, exposing systemic fragility from highly leveraged bets on AI infrastructure stocks [1].

China's AI industry is accelerating into multi-agent collaboration and edge-cloud integration, with breakthroughs like Alipay's AHA protocol, HIT's GUI Agent, and Bilibili's Updream Preview Platform. AI hardware demand is fueling memory chip growth—Puyan reported a 1929% YoY net profit surge in H1. Meanwhile, large-model monetization is shifting from 'selling reach' to 'selling outcomes.'

China's AI industry is rapidly transitioning from the technology validation phase into the commercialization phase: Baidu's Q2 AI revenue reached 12.5 billion RMB—over half of its core business revenue [5]; Alipay launched the AHA multi-agent cross-device interoperability protocol, shifting intelligent agents from a 'reach-selling' to a 'result-selling' paradigm [2]; meanwhile, foundational conceptual divergences—including the open-source document parsing tool AnyDoc and the 'internalized knowledge-first' thesis—are reshaping standards for evaluating model capabilities [3][4].

Anthropic has mandated statistical watermarking across all Claude outputs—a move criticized for overreaching the scope of the EU AI Act and negatively impacting creative use cases [1]; meanwhile, the DeepSeek Harness plugin ecosystem is experiencing explosive growth, with users transforming it into a 'cyber LEGO' platform supporting multi-agent orchestration and customizable UI themes [2].

DeepSeek Harness Introduces 'Everything-as-a-Plugin' Architecture · 0818-580

The AI industry is rapidly shifting from 'model capability contests' to 'engineering paradigm reconstruction' and 'end-to-end task validation'; Agent engineering, real-world scenario evaluation benchmarks, and knowledge infrastructure have become critical differentiators. Alibaba's newly released RealReplicaBench exposes that 13 mainstream models collectively fail e-commerce tasks [2], underscoring the paradigm shift: 'being able to answer questions' does not equate to 'being able to execute tasks.'

A paradigm shift has arrived in e-commerce AI Agent evaluation: Alibaba's Accio Work team launched RealReplicaBench—a rigorous, real-world task-flow benchmark—under which all 13 mainstream models failed to pass, signaling a decisive pivot from 'answering questions correctly' to 'getting things done' [1]. Concurrently, rising AI emotional dependency among minors has surfaced, underscoring the urgent need for collaborative governance across families, industry, and education [4].

Anthropic's Q2 revenue jumped ~1400% YoY; company eyes a record-breaking $2 trillion IPO in October. Meanwhile, Hugging Face reports Chinese open-source models now lead U.S. counterparts in parameter count, and AI agents are emerging as key open-model users.

AI-powered information filtering is shifting from manual subscription toward intelligent agent-driven curation, with RSS source tiering and user preference learning emerging as critical strategies against information overload. Meanwhile, Meta's 'Personal Superintelligence' initiative elevates the narrative of AI accessibility to new heights—though its real-world implementation remains constrained by historical deficits in public trust [1][2].

Indium phosphide substrate shortages surge amid AI compute arms race; Claude's SynthID-Text watermarking revealed—embedding invisible statistical traces via RNG seed manipulation; Anthropic's Panama Project cleared as fair use by U.S. court after AI firms bulk-purchase used books for training.

OpenAI v2, Flue 2, and Astro are converging on an 'agent-as-primitive' paradigm for next-gen AI. Meanwhile, 1:1 biometric matching (via Euclidean distance) proves critical against AI identity spoofing, and China-led open models like Qwen are reshaping the global foundation model landscape.

GLM-5.3 improves coding performance by 50% and detects 2,404 security vulnerabilities in two weeks; Anthropic targets $2T IPO valuation; Alibaba open-sources LongHorizon-Harness for long-horizon agent state management; NVIDIA cuts OpenAI datacenter guarantee to $120B, becomes SpaceX's sixth-largest shareholder.

Anthropic's quarterly revenue surges to $1.15 billion—up 14x year-over-year—with a potential $2 trillion IPO expected in October. Meanwhile, Claude launches invisible text watermarking to comply with EU AI Act, though detectability and accountability remain contested.