Xiaoliu BOT
中文EN

Briefs

Claude Fable 5 Disproves Jacobian Conjecture, Hugging Face Hit by AI Agent Attack, US-China AI Gap and Model Competition

This daily brief covers major AI developments: Claude Fable 5 disproving the long-standing Jacobian Conjecture, a sophisticated AI Agent attack on Hugging Face thwarted by an open-source model, and analysis of the US-China AI gap alongside releases like Kimi K3, Qwen3.8-Max…
Kimi K3 Pauses New Subscriptions, Qwen3.8-Max Preview Released, Basalt Labs Fake Model Exposed

This AI briefing covers major industry developments: Kimi halts new K3 subscriptions due to overwhelming demand, Alibaba launches the powerful Qwen3.8-Max preview model, and a social experiment by Basalt Labs reveals how easily a fake 'world's best' model can deceive the…
Claude Fable 5 Added to Subscriptions, Kimi K3 Capabilities Verified, AI Development Enters a New Stage

Catch up on the latest July 18 AI industry updates: Claude Fable 5 is permanently added to Anthropic subscription plans, Kimi K3 proves its leading performance as an open-source model, Meta and Anthropic negotiate a $10B compute deal, and AI-assisted development enters a new…
Kimi K3 Open-Source 3 Trillion Parameter Model Tops Programming, WAIC Establishes First AI International Body, Suno Data Leak Raises Concerns

This AI briefing covers major developments: Kimi's K3 model leads in programming tasks as the first open-source 3-trillion-parameter model; WAIC 2026 sees 29 nations form the first intergovernmental AI organization; Suno suffers a data breach revealing training data sources…
Kimi K3 Open-Source Model Released with Performance Nearing Top-Tier; Grok Build and Coding Tools Successively Open-Sourced or Reset Quotas

July 16 AI Brief: Kimi K3 open-sourced with near-top performance, Grok Build open-sourced, coding tools quota resets, Anthropic's Ode enterprise push, Thinking Machines' Inkling, OpenAI's hardware debut and GPT-Red.
Apple Intelligence Gains China Approval, DeepSeek Launches New Funding Round, Cutting-edge AI Regulation Framework Sparks Debate

Apple Intelligence receives official approval to operate in China, with Alibaba's Qwen AI to be integrated. DeepSeek reports $4-5B annual revenue, launches a new ~$74B funding round, and plans a Shanghai IPO. Google DeepMind's CEO proposes a new US-led AI safety review body…
IBM Plunge Reveals Corporate IT Budget Shift to AI, Tencent's Flagship MoE Model Runs on Single GPU, OpenAI-Anthropic Rivalry Intensifies

This AI briefing covers major industry shifts: IBM's stock crash signals corporate IT budgets pivoting to AI infrastructure; Tencent achieves a breakthrough by compressing its 295B MoE model to run on a single GPU; the OpenAI-Anthropic rivalry escalates with price cuts and…
AI Giants Compete for Users, Grok Build Privacy Concerns Heat Up, AI-Generated Content Enters Commercial Arena

Major AI companies like OpenAI, Anthropic, and xAI are aggressively rolling out user-friendly updates and feature expansions, turning intense competition into direct benefits for subscribers. However, xAI's Grok Build faces significant privacy backlash for allegedly uploading…
GPT-5.6 Sol Achieves Top Rankings Across Benchmarks, Noise Image Tests Expose Model Variances, AI Video Generation Advances to Precision Control

AI daily brief: GPT-5.6 Sol dominates benchmarks, noise probing shows model quirks, Grok 4.5 gains Musk's endorsement, GLM 5.2 offers best cost-performance, AI video generation achieves fine control, SpaceX Starship launch set, and agent infrastructure evolves with thinner…
Apple Sues OpenAI for Theft, GPT-5.6 Sol Raises Safety Concerns, AI Video Workflows Emerge

Apple sues OpenAI for systematic trade secret theft involving 400+ former Apple employees; GPT-5.6 Sol is praised as a daily assistant but Claude Fable excels in complex tasks; a Sol agent accidentally deletes user files, sparking safety debates; OpenAI launches ChatGPT Work…
GPT-5.6 Officially Released, Zuckerberg Returns to X Platform for Counter-Launch, MiniMax Secures Massive Funding

OpenAI launches GPT-5.6 with a revamped product suite. Meta counters with Muse Spark 1.1 via Zuckerberg's X return. Grok 4.5 leads benchmarks. MiniMax raises $2B, Tencent may lead a Manus buyback, Ollama secures $65M, and OpenAI's app CEO resigns.
GPT-5.6 and Grok 4.5 Launch Same Day, Claude Fable 5 Delayed, MIIT Warns of Claude Code Security Risk

OpenAI releases GPT-5.6 Sol/Terra/Luna on July 9, while SpaceXAI launches Grok 4.5. Anthropic delays Claude Fable 5 to July 12 and expands Cowork to cloud. Cloudflare introduces per-use billing for AI agents. China's MIIT warns of backdoor risks in Claude Code. Meta unveils…
Anthropic discovers model's internal 'thinking space', DeepSeek develops own inference chip, China plans to restrict overseas access to AI models

Anthropic reveals Claude's J-space thinking region; DeepSeek develops proprietary inference chip; China considers limiting overseas access to top AI models; plus Claude Code history, motion-anything open-source, and X platform updates.
Fable 5 vs Codex Practical Comparison, NVIDIA Free GLM 5.2, OpenClaw Reaches 100k Issues

Fable 5 outperforms Codex in complex coding tasks; NVIDIA offers free GLM 5.2 with strong front-end design; OpenClaw hits 100k issues amid controversy; plus Codex tool updates, blog migration, and more.
GPT-5.6 to Launch Next Week Targeting Fable Users, Doubao and Qianwen Remove Agent Features, Gemini 3.5 Pro Delayed

OpenAI plans GPT-5.6 launch next week targeting Fable 5 users; Doubao and Qianwen remove agent features ahead of China's AI anthropomorphism regulation; Gemini 3.5 Pro delayed to July 17; Fable 5 ecosystem expands; AI talent market shows strain; cognitive gap in AI capabilities…
Alibaba Bans Claude Completely, Triggering US-China AI Bilateral Decoupling; Meta and OpenAI Race to Release New Models

Alibaba bans Claude after backdoor discovery, signaling US-China AI decoupling. Meta's Watermelon rivals GPT-5.5; Codex usage up 6x. Google, Anthropic launch new products. Fable 5 jailbroken. OpenAI plans US govt stake.
OpenAI Proposes Giving Shares to Government for Universal Fund, Anthropic Attracts Top Talent, Palantir CEO Criticizes AI Business Model

OpenAI proposes 5% shares to US for universal AI fund; Anthropic hires top academics; Palantir CEO criticizes AI model; DeepSeek outage; SiliconFlow IPO; Unitree IPO; ArXiv redesign; Markov open-sources game dataset.
Claude Sonnet 5 Launch Sparks Cost Controversy, Fable 5 Returns Globally, Neuralink Achieves Dural Puncture Implant

Anthropic launches Claude Sonnet 5 amid token cost debate, lifts Fable 5 export ban; Neuralink performs first dural-puncture electrode implant; Google unveils fastest image model; Claude Science workstation released.
Claude Sonnet 5 Pricing Leaked, X Launches Official MCP Service, Anthropic Account Ban Controversy Escalates

Today's AI briefing covers Claude Sonnet 5's launch pricing at $2/$10 per Mtok, X's official MCP service enabling direct platform data access for agents, OpenAI's custom keyboard for Codex, Anthropic's escalating account bans with email tracking and proxy interception, Qwen's…
DeepSeek V4 Pricing Adjustments, GPT-5.6 Grayscale Testing Attracts Attention, AI Quality Inspection Encounters Bottlenecks

DeepSeek V4 pricing changes, GPT-5.6 grayscale testing, Claude bans, Tesla FSD v14 Lite, AI in IM, Jia Yangqing leaves NVIDIA, Ford AI QC struggles, open-source Figma Motion alternative.
Grok 4.5 Private Beta Begins, Performance Nears Opus; GPT-4.5 Officially Retired; Alibaba Releases Real-Time Interactive Model

Grok 4.5 private beta nears Opus; GPT-4.5 retired; Alibaba WanStreamer; Google Design Variations; Codex updates; Tokei v1.08; World Cup AI predictions.
GPT-5.6 Released in Three Tiers but Access Restricted; Anthropic Partially Unblocks Models; DeepSeek's Accelerator Framework and Funding Draw Attention

Today's AI news covers OpenAI's GPT-5.6 release in three tiers (Sol, Terra, Luna) with limited access due to US government request; Anthropic partially restores Mythos 5 access while Fable 5 awaits approval; DeepSeek's funding round and DSpark acceleration framework; OSWorld…
Apple Raises Prices Across the Board, GPT-5.6 Released in Tiers, DeepSeek Launches Massive Hiring Drive

Apple raises prices across Mac/iPad lines by up to 20% due to AI data center expansion costs. The Trump administration requires OpenAI to release GPT-5.6 in a tiered, per-client approval model. DeepSeek announces a massive hiring plan to double its team. UC Berkeley professor…
Anthropic Alleges Largest-Ever Distillation Attack by Alibaba, ByteDance Releases Seed-2.1-Pro, Qualcomm Acquires Modular

This daily AI brief covers significant industry developments: Anthropic's formal accusation against Alibaba's Qwen for a massive distillation attack on Claude; ByteDance's release of the coding-focused Doubao-Seed-2.1-Pro model; Qualcomm's $3.92B acquisition of AI…
OpenAI Unveils Self-Developed AI Chip Jalapeño, Anthropic Launches Slack Collaboration AI Claude Tag, ByteDance Updates Multiple AI Products

This daily AI brief covers major developments including OpenAI's first in-house AI chip Jalapeño, Anthropic's launch of Claude Tag for Slack collaboration, and a suite of updates from ByteDance's AI products. It also highlights Starlink's accelerated deployment in aviation and…
ByteDance Volcano Engine Unveils Full AI Upgrade, OpenAI Expands Cybersecurity Program, Baidu Open-Sources Long-Document OCR Model

ByteDance Volcano Engine unveils Seedance 2.5 & Seed 2.1 Pro; OpenAI launches GPT-5.5-Cyber & Codex Security; Baidu open-sources Unlimited-OCR. Also: Claude outage, Google Interactions API, Codex fix, talent moves, model delays.
WeChat Agent 'XiaoWei' Opens Beta, AI Browser Plugin Trojan Affects 900K Users, Anthropic Completes New Mythos Training

WeChat Agent XiaoWei beta exceeds expectations; AI browser plugin trojan affects 900K+ users; Anthropic finishes new Mythos training; Alibaba team-building signals stability; Sakana AI releases Fugu orchestration; Codex/Claude Code tools emerge; GPT-Image-2 generates…
DeepSeek Harness Group Launches Large-Scale Hiring, Mythos Breaches Pentagon Systems, SubQ 1.1 Small Enhances Attention Efficiency

DeepSeek Harness hiring, Mythos breaches Pentagon, SubQ 1.1 boosts efficiency, Cowart open-sourced, Doubao AI video, OpenClaw update, GPT-Image-2 design, Hermes Agent, icon generator.
DeepMind Core Scientist Joins Anthropic, Codex Launches Cross-Device Task Migration, WeChat AI Assistant Xiaowei Begins Beta Testing

DeepMind's AlphaFold lead John Jumper joins Anthropic; Codex introduces Handoff for cross-device task migration; WeChat AI assistant Xiaowei enters beta; plus tips for GPT Image 2 and Agent engineering practices.
Codex Launches Record & Replay, GPT-5.6 Pro Secretly Tested, Claude Code Adds Artifacts

Key AI updates: Codex Record & Replay, GPT-5.6 Pro test, Claude Code Artifacts, Fable 5 return, EverOS, Doubao voice, Grok on Databricks, Manus buyback.
Anthropic Export Control Controversy Continues, Google Core Talent Joins OpenAI, China Encourages AI Consumer Loans

Today's AI highlights: the Anthropic/Fable 5 export control saga escalates with White House, G7, and internal employee reactions; Noam Shazeer, co-founder of Google Gemini, joins OpenAI; DeepSeek launches its V4.1 vision model; China's eight ministries encourage consumer loans…
SpaceX Acquires Cursor and Launches Code Hosting Service, GLM-5.2 Released Open Source, Anthropic's Frontier Models Designated as National Security Assets

This daily AI brief covers major industry moves: SpaceX's $60B acquisition of Cursor and new Origin service, the open-source release of Zhipu's GLM-5.2 model, and the U.S. government restricting foreign access to Anthropic's advanced AI models. Additional highlights include…
SpaceX Acquires Cursor, DeepSeek Completes Special Structure Financing, Anthropic Faces Multiple Crises

SpaceX acquires Cursor for $60B; DeepSeek raises 50B yuan; Anthropic faces Pentagon shift, class action, model takedowns; Microsoft sued; Mistral rumor; ByteDance Seedance 2.0 Mini; Novo Nordisk hack; dev tools updates.
Anthropic's Multi-Front Turmoil, Loop Engineering Becomes New Consensus, AMD Launches Local AI Development Platform

Anthropic faces Fable 5 removal, KYC, deception tests. Loop Engineering gains traction. AMD launches $3,999 local AI platform. MiniMax M3 half-price. Zhipu GLM-5.2 open-sourced, stock +48%. SpaceX $2.11T. Nadella's Token Capital.
Alibaba Denies Rumors of Chief Scientist's Departure, Brazil's Open-Source Model Accused of Shelling, Anthropic Export Control Details Revealed

This AI briefing covers Alibaba's formal denial of its chief scientist's rumored departure, allegations that Brazil's Rio 3.5 model is a shell of Nex-N2, and conflicting narratives on Anthropic's Fable 5 export controls. It also details OpenRouter's multi-model Fusion…
U.S. Government Bans Foreign Nationals from Accessing Claude Fable 5; Alibaba Chief Scientist Zhou Jingren Resigns; Zhipu AI Releases GLM-5.2 Open-Source Model Amid Regulatory Trends

The U.S. government has imposed emergency export controls on Anthropic's Claude Fable 5 and Mythos 5 models, barring foreign nationals from access, including non-U.S. employees, disrupting numerous projects. Alibaba's chief scientist Zhou Jingren resigns shortly after a…
SpaceX Record IPO, Multiple AI Models Released Densely, OpenAI Acquisition Strengthens Execution Capabilities

SpaceX's record $75B IPO creates 4,400 new millionaires. MiniMax-M3 open-sourced with 428B parameters. Huawei Pangu revealed as China's first true LLM architecture. OpenAI acquires Ona for Codex execution layer. Tesla FSD 14.3.4 rolls out, Belgium becomes 5th European approval.
Anthropic Fable 5 Sparks Token Cost and Safety Debates, Alibaba Restructures AI Division, Google Open-Sources DiffusionGemma

Anthropic's Fable 5 model triggers token consumption surges and safety policy backlash; Alibaba overhauls its AI leadership and structure; Google open-sources DiffusionGemma, a text diffusion model challenging autoregressive dominance.
Anthropic's New Model Fable Imminent, Apple Siri AI Limited to iPhone 17 Pro, SpaceX Reveals Space AI Infrastructure Blueprint

Anthropic's new model Fable launches June 10 with enhanced safety; Apple's WWDC2026 Siri AI runs on-device only on iPhone 17 Pro; SpaceX details AI satellites, Terafab chip factory, and lunar mass driver; Xiaomi achieves 1000+ token/s on trillion-parameter model; OpenAI files…
WeChat AI Ecosystem Opens Mini-Program Access, OpenAI Prepares Biggest ChatGPT Overhaul, Notion Bans Anthropic Models

WeChat AI opens mini-program access; OpenAI plans ChatGPT overhaul; Notion disables Anthropic models. Also: Claude Design workflow, Kimi Work agent, ByteDance AI4S conflict, AI relay warning, AI video tools, Anthropic code, AI PPT.
Cursor Launches Design Mode Revolutionizing UI Design, Raycast Glaze Enables One-Sentence Mac App Development, Mainstream AI Model Capabilities Spark Debate

Cursor Design Mode, Raycast Glaze streamline UI design, Mac app dev. Opus leads in aesthetics, GPT in engineering; DingTalk letter, Grok Build updates, OpenAI super app rumors.
OpenAI Account Ban & Restore, Claude Mythos 5 Model Codename Surfaces, SpaceX Lands Google Compute Deal

OpenAI restores accounts after a system error caused mass bans; a Team membership exploit sparks a free-claim frenzy. Claude Mythos 5 model slug appears in API, Cowork usage doubles. SpaceX signs a massive compute contract with Google, while Codex and Cursor roll out UX…
ChatGPT Memory System Upgrades to Auto-Extract, Kimi Work Schedules 300 Agents, Starlink V3 Bandwidth Increases 100x

OpenAI upgrades ChatGPT memory with Dreaming architecture, boosting factual accuracy from 41.5% to 82.8%. Kimi Work shifts from coding to office tasks, orchestrating up to 300 agents. Starlink V3 achieves 100x bandwidth increase, orbit lowered to 350km. Supabase raises $500M at…
SpaceX sets record IPO valuation at $1.77 trillion, internet bot traffic surpasses humans for first time, Google open-sources Gemma 4 12B

SpaceX files record $1.77T IPO; internet bot traffic surpasses humans (57.5%); Google open-sources Gemma 4 12B; Evan You's VoidZero joins Cloudflare; Grok tops multiple benchmarks; Ideogram 4.0 open-sourced; VS Code vulnerability disclosed.
Codex Massive Updates Evolve Toward All-In-One Platform, Microsoft Launches RTX Spark Dev Machine, ByteDance Unveils Three New Releases

Catch up on the latest June 3 AI industry news: OpenAI Codex’s massive updates evolving into an all-in-one platform, Microsoft’s new RTX Spark dev machine, ByteDance’s three new releases, GPT-image-2 prompting tips, Stanford’s latest AI research and more key updates.
AI Companies Rush to IPO, WeChat Agent Launch Set for Q3, Doubao Pricing Sparks Debate

This daily brief covers major AI industry movements: Anthropic, MiniMax, and Zhipu AI file for IPOs; WeChat plans a deep integration of its Hunyuan model with a built-in Agent for Q3; ByteDance's Doubao AI announces a three-tier subscription model, sparking debate on…
NVIDIA Launches Three Major Products, MiniMax Unveils M3 Model, VAST Secures Nearly $200 Million in Funding

NVIDIA launches Nemotron 3 Ultra, RTX Spark, and Cosmos 3 at Computex 2026. MiniMax releases M3 with 1M context and sparse attention. VAST raises ~$200M, becoming a unicorn. Claude Opus 4.8 receives mixed reviews. Coze 3.0 enables local agent integration. Vibe Coding tools…
GPT-5.5 Leads in Coding, NVIDIA Enters PC Processor Market, Universal Agent Seen as Future OS

GPT-5.5 tops coding benchmark; NVIDIA enters PC processor market; DeepMind podcast on Gemini self-learning; Codex hits 5M users; GPT-Realtime 2.0 demo; Wharton professor joins OpenAI; Tianya relaunch; Universal Agent as future OS.
Codex Lands on Windows with Remote Control, GPT-5.5 Positioning and Efficiency Gains, Agent Token Optimization Methodology

OpenAI brings Codex Computer Use to Windows, enabling remote control via phone. GPT-5.5 boosts efficiency. Developers share Agent Token optimization using Skill+Script. Also: AI video creation workflows, GitHub trending social card skill, and iOS reader app.
Claude Opus 4.8 Released, Anthropic Valuation Surpasses OpenAI, AI Programming Agent Commercialization Booms

Anthropic releases Claude Opus 4.8, emphasizing honesty and launching Dynamic Workflows, as its valuation hits $965B, surpassing OpenAI. The AI coding agent market enters a commercial boom with Cognition's $10B+ funding and Cursor's report showing AI reshaping development…