868 亿！黄仁勋官宣收购 Hugging Face，AI人才的黄金赛道又变了？
sohu.comNvidia's $12.9B acquisition of Hugging Face analyzed as major shift in AI talent landscape; article explains that Nvidia is buying an open source community and ecosystem, not just a company - significant implications for transformers library users and developers.
Nvidia buys Hugging Face for $12.93B; OpenAI hack prompted CEO to sell
msn.comNvidia acquires Hugging Face for $12.93B following OpenAI testing environment hack incident; deal includes ~$11.9B to shareholders and up to $1B in employee equity incentives, expected to close by first half of 2027.
SWE-Serve揭示本地测试与实际服务的差距
news.qq.comSWE-Serve evaluates inference service changes with SGLang team assistance, testing 53 tasks including real-time service checks to identify gaps between local tests and production deployment.
从“单点优化”到“规模化提升” 百度智能云构建Agent时代推理基础设施
news.qq.comBaidu Smart Cloud and SGLang community meetup on building inference infrastructure for agents, covering multi-core adaptation, context caching, KV cache management.
I stopped paying for multiple AI coding tools and built one setup that can use Claude, GPT, Gemini, and local models
xda-developers.comThe author discusses using OpenCode as an open-source AI coding agent in the same category as other tools, building a setup that can use multiple models including local ones instead of paid subscriptions.
The story of how I solved a Kindle opening error with Claude Code
note.comUser shares experience solving a Kindle opening automation task using Claude Code, demonstrating practical AI coding tool usage.
Anthropic Launches Official Cloud Sessions for Claude Code with $250 Free Benefit
note.comOn September 23, 2026, Anthropic officially released Claude Code's cloud sessions feature for users who close their PCs at night, offering a $250 free benefit to existing subscribers until October 8.
智谱GLM-5.3-Flash日调用量增长73%
yicai.comOpenRouter daily leaderboard shows Zhipu AI's GLM-5.3-Flash reached 4.36 trillion tokens in calls, a 73% increase over the previous period, ranking first among platform models.
DeepSeek Harness桌面版现身，尚未正式官宣
msn.comDeepSeek Harness desktop client download links have appeared and can run normally after installation. The AI lab has not officially announced it yet, but the client provides a complete graphical operation interface.
China's DeepSeek annualised revenue hits $1 billion, The Information reports
msn.comDeepSeek's annualized revenue run rate has reached $1 billion, more than doubling from the previous year. The AI startup is scaling its API business with significant price increases driving growth.
Everything new in AI | DeepSeek returns, Veo 3 beaten & more
msn.comAI news roundup covering DeepSeek's return, Amazon's secret AI models, Google's breakthrough, OpenAI updates, and competition facing Veo 3.
When ChatGPT has too many line breaks and is hard to read: 3 steps to isolate the issue before rewriting the text
note.comTechnical discussion about ChatGPT's formatting issues with excessive line breaks and how to isolate the problem. Published 2026-09-19.
ChatGPTとClaudeが、最も煩わしくて時間の無駄になる習慣を改めた | The Star 2️⃣ 世界中の企業や政府機関のウェブサイトに侵入した後、ChatGPTの開発元である……
note.comArticle discussing ChatGPT and Claude's habits that waste time, with mentions of security incidents involving OpenAI. Published 2026-09-19.
Claude Sonnet 5.5开启灰度测试；MiniMax新文本模型上线编程平台；小米开源MiMo新版模型/
sohu.comMiniMax released M3.1-Flash-Preview model and launched it on MiniMax Code platform, positioning as a multimodal programming model supporting 1M token context with adjustable inference depth.
A new AI patent by Nvidia may solve developer woes surrounding game optimization
notebookcheck.netNVIDIA filed a new patent allowing developers to consult AI models like chatbots for diagnostics of existing game optimization issues, addressing developer challenges in the gaming industry.
Nvidia’s CEO Expects AI-Powered Cybersecurity to Boom. These 2 Companies Are Poised to Get a Big Boost From the Trend.
finance.yahoo.comNvidia CEO Jensen Huang stated that cybersecurity would be the next major area of growth for AI, highlighting Nvidia's strategic focus on emerging AI applications.
MaaseAI Introduces Security AI Model for Enterprise AI Protection and Governance
manilatimes.netMaaseAI launches a new Security AI Model framework emphasizing data classification, authorized access, encryption, and auditable flows for enterprise AI governance.
Pacific media needs stronger AI strategy, research finds
postcourier.com.pgResearch indicates Pacific media organizations using artificial intelligence urgently need strategic policy and technical support to effectively implement AI solutions.
Letitia James, fellow AGs push Congress on AI safety
msn.comMore than two dozen attorneys general are urging congressional leaders to establish federal safety rules for AI, highlighting growing industry and regulatory concerns.
Warner pushing new AI safety regulation legislation
cbs19news.comVirginia Sen. Mark Warner is calling for new safety rules around artificial intelligence, citing recent incidents that demonstrate the need for regulatory intervention.