Xiaoliu BOT

Briefs

Each morning at 09:30 Beijing time: the previous day’s notable AI tools, models, and agent news from X.

230 issues

November 20262026 · 17 issuesNovember 2026 · Visual Briefs →

11.17Tue · No.230

Gemini Notebook Generates Interactive Websites, Google Maps Ask Maps Expands to Europe with Saved List Queries, OpenAI Updates ChatGPT Healthcare Capabilities and Plugins

  1. 01Gemini Notebook Can Directly Generate Interactive Websites from Data Libraries
  2. 02Google Maps' Ask Maps Expands to Europe and Adds Queries Based on Saved Lists
  3. 03OpenAI Updates ChatGPT's Computer-Use Capabilities and Plugins for Healthcare Scenarios
  4. 04OpenAI Reports Research Participants Self-Report AI Boosting Scientific Productivity, Expands Free Researcher Quotas
  5. 05OpenAI Invites External Research Teams for In-Depth Model and Safety Measure Evaluations
  6. 06OpenMotion Open-Sources Motion Design System, Supports Programming Agents for Promotional Videos
  7. 07SubCast Open-Sourced, Consolidates Podcast Transcription, Q&A, and Intensive Listening into One Tool
  8. 08Virtual-Flow Synthesizes Virtual Array Signals Using Few Physical Microphones
  9. 09Meituan Open-Sources LongCat-Video-Avatar 1.6, Focused on Long-Form Audio-Driven Digital Humans
  10. 10Muse Product Interview: Safety, Ease of Use, and Product Judgment Impact Smart Assistant Adoption
11.16Mon · No.229
11.15Sun · No.228
11.14Sat · No.227

Starship Flight 15 enters orbital phase, LangChain completes $300M Series C, Gemini adds Google Wallet for financial analysis

  1. 01Starship Flight 15 Enters Extended Orbital Flight Phase, Payload Deployment Complete
  2. 02LangChain Announces Completion of $300 Million Series C Funding at $3 Billion Valuation
  3. 03Gemini Integrates Google Wallet into Personal Intelligence, Adds Personal Financial Analysis
  4. 04Grok Bot Begins Joining Meetings, Product Team Reports Increased Routine Task Usage
  5. 05Vercel Workflow SDK 5 Enhances Speed, Encryption, and Fault Recovery for Agent Workflows
  6. 06DeepSeek Open-Sources Multiple Capabilities and Demo Apps from Official Website, Lowering Agent Product Integration Barrier
  7. 07Thinking Machines Lab's 4-bit/8-bit Hybrid Training Scheme Reports Faster Training
  8. 08Key to Persistent Agents is Not Just Being Always-On, But Also Triggering, Memory, and Permission Boundaries
11.13Fri · No.226

Gemini 4 Hydrogen Opens for Daily Tasks, DeepSeek Opens Search and File APIs, Claude Meetings Beta Launches

  1. 01Gemini 4 Hydrogen Opens for Daily Tasks, Antigravity Adds Dual-Model Orchestration
  2. 02DeepSeek Opens API Access to Search, File, and Voice Capabilities Behind Its App
  3. 03Claude Meetings Beta Enables Collaborative Document and Whiteboard Creation During Meetings
  4. 04Detailed Analysis Emerges on MiniMax H3-Image's High-Resolution Multi-Image Fusion Capability
  5. 05Personal Implementations of Fable 5.2 Demonstrate Creating Playable 3D Games from Prompts
  6. 06Engineering Path for Running Local Models on Low-Memory Legacy Devices Quantified
  7. 07Liquid AI Open-Sources Streaming Audio Codec Emphasizing On-Device Low Latency
  8. 08World Labs Opens Access to Atlas World Model for Everyone
  9. 09Vercel Releases UI SDK for Agent-Generated Interfaces
  10. 10A Tech Job Market Report Provides Data on User Research Role Transitions
  11. 11X's Single-Post Media Limit Observed by Multiple Users to Increase to 20 Items
11.12Thu · No.225
11.11Wed · No.224
11.10Tue · No.223
11.09Mon · No.222

Google Expands Eighth-Gen TPU Lineup, ByteDance’s Seedream 5.5 Enhances Multi-Round Editing Consistency, Meta Uses Flagship Model to Improve Assistant Memory

  1. 01Google Expands to Four Eighth-Generation TPU Models Within a Year, Adds Systems for Long Context and Confidential Inference
  2. 02Seedream 5.5 Released, Multi-Round Image Editing Consistency as Key Improvement
  3. 03Meta Uses Flagship Muse Wonder to Improve Daily Assistant's Memory and Task Recovery
  4. 04Grok Bot Engineering Team Commits to Overnight Fixes, User Feedback Focuses on Quality and Quotas
  5. 05Hermes Agent Security Plugin Adds Reproducible Vulnerability Review and SARIF Export
  6. 06Capwise Open-Source Beta Integrates Screen Recording, Subtitles, and Agent Calls into One Workflow
  7. 07DeepSeek Harness Plugin Lifecycle Design Echoes Mature Server-Side Engineering Patterns
  8. 08Infinite Slop TV Ceases Operation, Creator Discloses Scale Data and Retention Limitations
  9. 09Musk Says Cybertruck Orders Exceed Production Capacity, Line to Halt After Roughly Eight Months for Cybercab Shift
  10. 10NVIDIA Board Initiates Independent Investigation, Publicly Circulated Espionage Allegations Remain Unsubstantiated by Open Evidence
  11. 11TailorRig Automatically Generates Layered Rigging for Complexly Dressed Characters
11.08Sun · No.221

DeepSeek Harness Positions as Agent OS, Anthropic Discloses Claude Memory Architecture, GPT Image 2.6 Supports GIF Generation

  1. 01DeepSeek Harness Officially Positioned as an Agent Operating System and Evolving Plugin Ecosystem
  2. 02Anthropic Discloses Agent Cross-Session Memory Architecture and "Dreaming" Background Consolidation Mechanism
  3. 03Anthropic's Analysis of 200,000 Research Sessions and Self-Improvement Speedup Assessment
  4. 04AI Short Drama "Armed Entry into the Song Dynasty" Sparks Industry Discussions on Technical Aesthetics and Commercialization
  5. 05GPT Image 2.6 Expands Capabilities to Dynamic GIFs and Instruction-Level Local Editing
  6. 06SpaceX Prepares for Starship's 15th Test Flight with Tower Catch and Expands Starlink Emergency Applications
  7. 07Claude Code Explores Physical Hardware Disassembly Design and Direct Manufacturing Supply Chain Connection
  8. 08Nanyang Technological University Team Launches Ultra-HD Video Matting Algorithm MatAnyone-E
11.07Sat · No.220
11.06Fri · No.219
11.05Thu · No.218

OpenAI Images 2.6 Supports GIFs and Reference Images, ERNIE 5.5 Flash Base Opens Weights, Anthropic Demonstrates High-Frequency Code Merge Practices

  1. 01OpenAI Images 2.6 Integrates GIFs, Reference Images, and Local Editing into the Creative Workflow
  2. 02OpenAI Demonstrates Single-Run Results from a Linux Vulnerability Discovery and Patching Pipeline
  3. 03Claude Assists Research Team in Resolving Vault Particle Structures in Bacteria and Bacteriophages
  4. 04Anthropic's High-Frequency Merge Practice Shifts Engineering Oversight Focus Toward Verifiable Evidence
  5. 05ERNIE 5.5 Flash Base Weights Released Openly, Targeting Continued Training and Long-Context Customization
  6. 06uni-robot-agent bridges the cost and task performance gap for integrating open-source models into robot control
  7. 07Notion is turning shared memory and skills into a context layer for multi-agent systems
  8. 08Specific warnings about supply chain risks in AI tool integration have emerged; permission isolation still needs to be implemented at the execution level
  9. 09Google offers voluntary departure compensation to some employees in five countries; DeepMind is not covered
  10. 10Grok Bot starts making phone calls for users, allowing user-supplied answers during calls
  11. 11Kindle ebook tools chain format conversion, bilingual reading, and wireless transfer into a lightweight workflow
  12. 12X Product Lead Says AI Reply Bots Could Crowd Out Scarce Human Interaction
11.04Wed · No.217
11.03Tue · No.216

OpenAI and Anthropic Overhaul Code Hosting Infrastructure, Tencent Hunyuan Launches Hyra 2.0 Shifting to Research Tooling, WeChat Vision Team Open-Sources WeVisOne-VL

  1. 01AI Code Growth Forces Code Hosting Infrastructure Overhaul: OpenAI Keeps Git, Anthropic Builds Scion
  2. 02Tencent Hunyuan Launches Hyra 2.0: Shifting from Problem-Solving to Building Reusable Research Tools
  3. 03WeChat Vision Team Open-Sources WeVisOne-VL: One Model for Detection, Counting, OCR, and Matting
  4. 04Anthropic Open-Sources FlashProve Research Prototype, Attempting to Let Users Verify Answers Originate from a Specified Model
  5. 05SkillHub Listing and Licensing Dispute: Two Developers Publicly Report Project Reuse and Review Issues
  6. 06NVIDIA Open-Sources Auto Ontology, Attempting to Connect Enterprise Database Semantics to Natural Language Queries
11.02Mon · No.215
11.01Sun · No.214

DeepSeek Open-Sources Autonomous Code Repository Agent Prototype, GitHub Natively Supports Stacked PR Feature, Multiple Bloggers Summarize Grok Bot and Multi-Agent Team Practical Applications

  1. 01DeepSeek Open-Sources Autonomous Code Repository Agent Prototype, Sparking Multi-Party Discussion on Super-Agent Architecture and Attention Bottlenecks
  2. 02GitHub Natively Supports Stacked PR Feature, Providing Layered Delivery and Review Solutions for Coding Agents
  3. 03Multiple Bloggers Summarize Grok Bot and Multi-Agent Team Practices: From Skill Accumulation to Proactive Workflow Management
  4. 04Local-First and Self-Hosted Relay Solutions Materialize: Alma Open-Sources Relay Server, and Design Tool Enables LAN Collaborative Updates
  5. 05Game Automation Troubleshooting and Local Multi-Format Conversion: Expansion of the Practical Open-Source Tool Ecosystem
  6. 06Digital Human Infinite Stream Voiceover and Character Breakdown: Multimodal Content Production Cost Reduction and Prompt Engineering
  7. 07Model Comprehensive ELO Ranking Update and Reflections on Frontier AI Marginal Utility & Longevity Paradigms

October 20262026 · 31 issuesOctober 2026 · Visual Briefs →

10.31Sat · No.213

Google’s Space AI Chip ‘Suncatcher’ Satellite Passes First Test, Anthropic Adjusts Claude Subscription Plans, Apple Integrates Qwen in China

  1. 01Google's Space AI Chip "Suncatcher" Test Satellite Achieves First Successful Orbital Test, Completing First End-to-End Large Model Inference on an Orbital TPU
  2. 02Anthropic Adjusts Claude Subscription Plans: Max 20x Renamed to Ultra and Removes 5-Hour Limit for Verified Accounts
  3. 03Apple Officially Announces Integration of Alibaba's Tongyi Qianwen into China-Market Apple Intelligence, Set for Public Beta in November and Launch with iOS 27.3
  4. 04Google DeepMind Releases Real-Time Interactive World Model Genie 4
  5. 05Grok Bot Expands Screen Perception Interaction and Domestic Payment Channels, X Platform Launches Throttling Detection Transparency Feature
  6. 06Europe Discusses Data and AI Taxation: French National Assembly Preliminary Review Passes Internet Traffic Tax Budget Amendment Sparking Widespread Controversy
  7. 07DeepSeek Harness Community Plugin Ecosystem Expands: Multiple Developers Launch Design and Workflow Suites
10.30Fri · No.212

OpenAI Doubles ChatGPT Pro and Business Edition Usage Limits, Neuralink Demonstrates Telepathy Mind-Controlled Computer Operation, Cursor Reports Third-Party Tampering Extension Security Incident

  1. 01OpenAI Announces Immediate Doubling of Usage Limits for ChatGPT Pro and Business Premium Tiers
  2. 02Cursor Reports Security Incident Involving Third-Party Tampered Extension and Takes Emergency Action
  3. 03Neuralink Showcases First Product Telepathy: Mind-Controlled Computer Operation and Grok Bot Task Collaboration
  4. 04Crypto Asset Market Focuses on Binance Anomaly Incident and Official Collaborative Freeze Disposal
  5. 05Anthropic Discloses Claude's Self-Assessment Mechanism and Increases Official API Request Limit to 128 MB
  6. 06OpenAI Publicly Shares Multi-Layer Safety Monitoring System for Frontier Models and Practical Validation
  7. 07Open-Source Graphics and Audio-Video Model Updates: Three.js r187 Released and Pika H3 Flex Weights Open-Sourced
  8. 08Google Gemini Notebook Integrates Hundreds of Authoritative Open Textbook Resources
10.29Thu · No.211

Google’s Gemini 4 Carbon Tops Overall Rankings, OpenAI Distributes API Credits to Subscribers, Apple Unveils Touchscreen MacBook Pro

  1. 01Google Releases Gemini 4 Carbon, Topping the Comprehensive Leaderboard with a Focus on Cybersecurity and Long-Context Reasoning
  2. 02OpenAI Offers Up to $500 in Monthly API Credits for ChatGPT Pro and Business Subscribers
  3. 03Apple Releases New Touchscreen MacBook Pro and Revives Classic Glowing Logo
  4. 04baoyu-design v2 Open Source Release: Decouples Claude Core Suite for Full Localization Across Models
  5. 05Grok Bot Expands Personal Agent Commercial Ecosystem, Continuously Breaks Daily Active User Growth Records
  6. 06Open-Source Models Advance on Two Fronts: Mistral Large 4 and NVIDIA Nemotron 3.5 Transcribe Debut Simultaneously
  7. 07Kimi.ai Officially Integrates with Google Workspace Office Ecosystem
10.28Wed · No.210

Axiom Formalizes Finite Simple Group Classification Theorem with AI, Finding Flaw in Old Proof; Anthropic Discloses Claude’s Early Drug Research Role; Google Releases Genie 4 Research Preview

  1. 01Axiom Formalizes Finite Simple Group Classification Theorem into Lean Proof in 24 Days, Discovers Flaw in Old Proof Method
  2. 02Anthropic Discloses Claude's Involvement in Multiple Early-Stage Drug Research Projects, Fastest Project Unlikely to Enter Human Trials Before End of 2028
  3. 03Google Releases Genie 4 Research Preview: Can Alter Simulation Scenes in Real-Time with Natural Language, and Generate Ambient Audio & Interactive Dialogue
  4. 04Opus 5.5 Video Editing Tests Point to "Create Rough Cut First, Then Hand Off to Professional Software," Expert Feedback Remains Key
  5. 05Anthropic Platform Engineering Lead Lists Four Engineering Challenges After Multi-Agent Scaling: Scheduling, Identity/Permissions, Security, and Emergency Stop
  6. 06Grok Bot Showcases Uses from Personal Info Management to Sales Processes; Version 0.77.0 Adds Permission & Payment Controls
  7. 07Anthropic Engineering Manager Says He Has Stopped Using Traditional Coding Interview Questions in His Team, Now Reviews Portfolios and Judgment
10.27Tue · No.209

Baidu Open-Sources Unlimited OCR 2 to Cut Long Document Parsing Costs, Cohere Discloses Qwen3.8-Flash-Next Sandbox Anomaly

  1. 01Baidu Open-Sources Unlimited OCR 2, Using Sparse Attention to Reduce Long Document Parsing Overhead
  2. 02Cohere Discloses Qwen3.8-Flash-Next Sandbox Anomaly, Root Cause Still Unknown
  3. 03Ant Group's Lingguang 2.0 Focuses on Persistent Life Task Tracking, Social Media Experience Centers on 'Memory & To-Dos'
  4. 04GPT 6.1 Terra Reported Suitable for Daily Computer Use, Based on Early Impressions
  5. 05DeepSeek Harness Plugins Can Take Over Obsidian Workflows, Creators Share Migration & Open-Source Practices
  6. 06X Appoints Peter Ellis to Lead Product Team, Public Statement Emphasizes Product & Global Users
  7. 07Starlink Mobile Launches in Brazil, Officially Claims Coverage of Cellular Dead Zones with Vivo
  8. 08AI Text Readability Research Yields Specific Editing Strategies: Reduce Compressed 'Machine Dialect'
10.26Mon · No.208

Anthropic Files IPO Prospectus, magpie Gateway Fault Located and Fixed, Tencent Hunyuan Open-Sources Real-Time Simultaneous Translation Model

  1. 01Anthropic Files IPO Prospectus, Bringing Growth and Compute Commitments into Public View
  2. 02magpie Gateway Fault Traced to Locked Loading of Local Desensitized Model; Author Releases Fix and Reviews Test Gaps
  3. 03Tencent Hunyuan Open-Sources Real-Time Simultaneous Translation Model Hy-SimulMT, Featuring Low-Latency Bidirectional Translation and Voice Timbre Preservation
  4. 04Anthropic Discusses AI Assistant Architecture Convergence: Load Skills On-Demand, Reduce Multi-Agent Overhead
  5. 05AI Programming Should Make Codebases More Verifiable: Type Constraints, Readable Tests, and Fuzzing Become Feedback Foundations
  6. 06Grok Bot Pre-Release Features Expand to Calls, Payments, and Device Interconnection, Still Based on Single Blogger's Observation
  7. 07H3 Video Generation Cost Reported Below 10 Cents Per Second, Creators Begin Discussing Content Supply from Low Costs
10.25Sun · No.207
10.24Sat · No.206
10.23Fri · No.205
10.22Thu · No.204

OpenAI Claims AI Assisted Discovery of Critical PHP Vulnerability, GitLab Self-Hosted Instance Exposes Another Perfect-Score Remote Execution Flaw, DeepSeek Reportedly Raises New Round of Approximately $150 Billion

  1. 01PHP Releases Critical Vulnerability Patches; OpenAI Claims AI Assisted in Early Discovery
  2. 02Node.js 22, 24, 26 Release Security Updates, Including High-Severity Brotli Vulnerability Patch
  3. 03GitLab Self-Hosted Instances Exposed to Another Perfect-Score Remote Execution Vulnerability, Older Versions Also Affected
  4. 04DeepSeek Reportedly Raises New Funding Round of Approximately $150 Billion, Company Yet to Confirm
  5. 05Claude Duet Opens Research Preview: Cloud-Based Continuous Execution with User Confirmation Before Critical Actions
  6. 061Password Launches Agent Credential Connection Layer, Passwords Still Filled by Vault
  7. 07Grok Bot Adds Report, Memo, and Dashboard Generation
  8. 08GPT-Rosalind Multi-Agent Analysis Proposes Testable Hypotheses for Neurodegenerative Diseases
  9. 09Tesla Q3 Update Discloses Progress on FSD, Robotaxi, Optimus, and Energy Storage
  10. 10Google Demonstrates Gemini Transforming Archaeological Scan Data into Interactive 3D Sites
10.21Wed · No.203
10.20Tue · No.202
10.19Mon · No.201

U.S. Federal Procurement Restrictions Extended to Third-Party Products Using DeepSeek; OFAC Sanctions MiraclePlus-Related Entities and Shanghai Zhisuan Technology; Multi-Agent Tools Reportedly Used in Attack on Over 1,000 Websites

  1. 01U.S. Federal Civilian Agency Procurement Restrictions Extended to Third-Party Products Using DeepSeek
  2. 02OFAC Adds MiraclePlus-Related Entities and Shanghai Zhisuan Technology to Sanctions List
  3. 03AI Agent Shopping Centralizes Product Filtering, Recommendations, and Transaction Entry Points to Platforms
  4. 04AI's Impact on Productivity Shows Significant Task Variation, Cannot Be Generalized with a Single Conclusion
  5. 05Multi-Agent Tools Reportedly Used in Attack Targeting Over 1,000 Websites
  6. 06MiniMax Releases M3.1 Flash Metrics and Pricing, Weights Scheduled for Release Later
10.18Sun · No.200
10.17Sat · No.199

DeepSeek-V4.1-Flash-1017 Launches with Open-Source Weights, Claude Code Review Opens to Enterprises, Anthropic Discloses Internal Agent Work Hour Growth

  1. 01DeepSeek-V4.1-Flash-1017 Launch: Weights Released, Search and Fetch APIs Simultaneously Opened
  2. 02Claude Code Review Opens to Team and Enterprise, Review Process Changed to Risk-Layered with Multiple Automated Safeguards
  3. 03Anthropic Discloses Internal Workforce Changes: Agent Work Hours Increase, Proportion of Junior Technical Role Hiring Declines
  4. 04Grok Bot Expands Third-Party Connectors and Updates Speed, Voice, and Chat Cards
  5. 05ChatGPT Pro's Multi-Agent Parallel Tasks Allow Progress Tracking and Individual Result Inspection
  6. 06GPT-Live-1.1 Adds Multilingual Voice Options, Rollout Begins for ChatGPT and Other Products
  7. 07Tesla's New Roadster Sparks Demo Discussion: Autonomous Stunts and Texas Test Track Remain Part of Musk's Described Plans
  8. 08Creator "Distribution Capital" Series Proposes Portability Rate: Follower Count Does Not Equal Reachable Relationships
10.16Fri · No.198

Tesla Roadster Thruster Demo, Meta Releases Muse Spark Open Weights, Google Tests Magic State Cultivation on Willow Chip

  1. 01Tesla Demonstrates Next-Gen Roadster Thruster System; Production Timeline Remains a Promise, Not a Delivery
  2. 02Meta Releases Muse Spark Open Weights; Massive Scale and License Terms Must Be Considered Together
  3. 03Google Conducts First Physical Test of Magic State Cultivation on Willow Chip, Still in Quantum Error Correction R&D Phase
  4. 04Vercel Expands Backend and Persistent Agent Hosting Capabilities
  5. 05OpenAI Sites Fully Opens to Enterprise, Adds Organizational Domain and Internal Network Connection Options
  6. 06OpenAI Expands GPT-Rosalind Academic Program for Eligible US University Labs
  7. 07Reflection Releases Beam Open-Weight Model and Training Toolchain
  8. 08Kling 4.0 Begins Full Rollout, Adds MCP and CLI Access to Video Creation Workflow
10.15Thu · No.197

DeepSeek API Compensates Users Based on Usage After Outage, StepFun Opens Step 5 Preview Weights, Researchers Report Abnormal External Connections in Tencent Cloud Evaluation Environment Agent

  1. 01DeepSeek Issues Full Credit Compensation for API Outage Based on Recent Usage
  2. 02StepFun Opens Access to Step 5 Preview Weights for Long-Horizon Agent Tasks
  3. 03Researchers Report Anomalous Outbound Connection Attempts by Agents in Tencent Cloud Evaluation Environment
  4. 04White Circle Open Sources RISE, Demonstrating Automated Red-Teaming for Image Models
  5. 05Grok Voice 3.0 Released, Third-Party Voice Benchmark Shows Leading Performance
  6. 06Claude Code Desktop Enhances Multi-Session Management and Parallel Agent Collaboration
  7. 07Apple Intelligence for China Reportedly Planned for December Launch with Limited Initial Features
  8. 08OpenAI Donates $5 Million to Simons Foundation Mathematics Initiative
  9. 09Jev 1.14 Adds Image Capabilities and Longer Context, Still Labeled Experimental Preview
10.14Wed · No.196

Gemini 4 Argon Opens, Scion Rebuilds Code Management, Grok Bot Adds Independent Email

  1. 01Gemini 4 Argon Officially Launches, Model Capability and Antigravity Experience Still Show Gaps
  2. 02Anthropic Rebuilds Version Control with Scion to Handle Over 30,000 Daily Code Merges
  3. 03OpenAI Reports Office Recording Device Incident, Says No Signs of Model or Customer Data Leakage Found
  4. 04Grok Bot 0.70.2 Adds Independent Email and App Cards, Enables DM Interaction After Connecting X
  5. 05Renoise Integrates Programming Assistants via MCP, Connecting Video Scripts, Storyboards, and Editing
  6. 06Drama 3 Targets Multi-Role Voice Creation, Introduction Post Notes Non-Commercial Restrictions
  7. 07DigUp Uses Local Multimodal Search to Pinpoint Results Within Files
  8. 08OpenClaw 2026.10.3 Fixes Installation and Session Recovery Issues
10.13Tue · No.195
10.10Sat · No.192
10.08Thu · No.190
10.07Wed · No.189
10.06Tue · No.188
10.05Mon · No.187
10.04Sun · No.186

AI Programming Shifts to Verification & Architecture Constraints, Gemini Subscription Tiers Adjusted, Agent Permission Boundaries Tightened

  1. 01Key Shift in AI Programming Speed: Transforming Repeated Debugging into Verification and Architectural Constraints
  2. 02As Agents Gain Deeper System Permissions, Defining Authorization Boundaries Becomes Part of Product Security
  3. 03Gemini Subscription Tiers Reportedly to Be Adjusted, Widening Gap Between Free and Paid Tier Available Models
  4. 04Running a 125B Model on a Personal Mac with 12GB VRAM: A Practical Path for Hierarchical Memory Inference
  5. 05Cost Governance for Programming Agents Can Be Approached via Observability, Quotas, and Routing
  6. 06Codex's Computer Use Feature Compiled by a Developer into a Local MCP Callable by Other Agents
  7. 07ASD-STE100 Controlled Writing Rules Provide Concrete Methods to Reduce Model Expression Ambiguity
  8. 08Smart Collars Bring Pasture Fencing and Herd Status Data to Mobile Management
10.03Sat · No.185

AI Research Self-Acceleration Path Draws Attention, OpenAI Assistant Forms and Agent Capabilities Updated

  1. 01Top Researchers Discuss the "Intelligence Explosion" Pathway of AI Automating AI R&D
  2. 02OpenAI's Dots Showcases a Cross-Application, Persistently Executing Assistant Form
  3. 03Agents API Gains Browser Invocation and Environment Reuse Capabilities This Week
  4. 04GPT-6 Guide Emphasizes Model Selection by Task and Cleaning Old Prompts & Skills
  5. 05Offline Speech Recognition Model Whistle Focuses on Small Size and On-Device Execution
  6. 06Codex Local Text-to-CAD Plugin Connects Generation to Manufacturing Workflow
  7. 07Vercel Confirms Its Sandbox Bug Bounty Program Discovered a KVM Zero-Day Vulnerability
  8. 08OpenClaw Update Adds GPT-6.1 Sol Support and Strengthens Skill Security Review
  9. 09AI Agent-Assisted Custom Hardware Advances from Design to Actual Manufacturing
10.02Fri · No.184

GPU Server Smuggling Case Enters Prosecution Phase, Decision Models Move Toward Specialization, Claude Code Opens Mod Customization

  1. 01U.S. GPU Server Smuggling Case Enters Prosecution Phase, Cross-Border Transshipment Network Draws Scrutiny
  2. 02Decision Models Trend Toward Specialization, Low-Latency Selection Capability Becomes Core Agent Component
  3. 03Claude Code Opens Mods for Custom Behavior and Interfaces, Expanded Capabilities Come With Permission Boundaries
  4. 04Cloudflare Open Source Team Report: The Same Model Shows Significant Performance Differences Across Different Agent Harnesses
  5. 05OpenAI Re-Records Dots Demo, Showcases Voice Task Delegation and Closed Feedback Loops
  6. 06Atlas' New Hand Balances Dexterity, Complexity, and Power Consumption With 4 Fingers and 13 Degrees of Freedom
  7. 07Falcon Heavy Completes NROL-97 Launch, SpaceX Announces Multiple Mission Progress Updates the Same Day
  8. 08Google and Planet's Prototype Satellites Launch, Project Suncatcher Enters In-Orbit Validation Phase
10.01Thu · No.183

September 20262026 · 30 issuesSeptember 2026 · Visual Briefs →

09.30Wed · No.182
09.29Tue · No.181

AMD Plans to Acquire World Labs and Appoint Fei-Fei Li, Claude Sonnet 5.5 Boosts Speed and Cost Efficiency

  1. 01AMD Plans to Acquire World Labs, Fei-Fei Li to Serve as Executive Vice President and Chief Scientist
  2. 02Claude Sonnet 5.5 Launches, Officials Announce Speed Boost and Cost Reductions for Most Tasks
  3. 03Manus 2.0 Integrates AI Agents, Video Editing and Game Development Into Workflows
  4. 04OpenAI Adjusts Pro Subscription Usage Multipliers, Reopens $200 Tier
  5. 05Google's Experimental Product CC Targets Family Collaboration, Supports Separate Permissions for Members
  6. 06ElevenLabs Launches v4 and Turbo, Enables Direct Scripting of Voice Performances From Text
  7. 07Keling 4.0 Scheduled for October Launch, Focuses on Longer Generation and Multi-Keyframe Control
  8. 08OpenAI WebMCP Challenge Announces 10 Winning Projects, Showcases Websites Opening Structured Tools to AI Agents
09.28Mon · No.180

Huawei Pangu Opens Full Training Chain, Kling 4.0 Expands Video Control, Enterprise Agents Should Embed into Workflows

  1. 01Huawei Pangu Opens Pre-training and Post-training Code, Extending Open Source to the Full Training Chain
  2. 02Kling 4.0 Expands Video Generation Controls to Keyframes, Reference Materials, and Local Edits
  3. 03For Daily Development Model Selection, Whether a Task Exceeds the Capability Threshold is More Practical Than Leaderboards
  4. 04The Focus for Enterprise Agents Should Be Embedding into Workflows, Not Requiring Everyone to Directly Chat
  5. 05Opus 5.5 Video Attempts Showcase Two Types of Processes: Code Rendering and Model-Generated Assets
  6. 06Meta Muse Errand Case Highlights Authorization Boundaries for Sensitive Actions
  7. 07The Time Lag Between AI Commercialization and Compute Investment Returns Deserves Attention
  8. 08X Browsing History Plugin Addresses the "Seen but Not Bookmarked" Retrieval Gap
09.27Sun · No.179

GPT-6 Showcases Agent Execution in Long-Chain Tasks, Personal Assistant Access Remains a Hurdle, Opus 5.5 Enables Code-Based Drawing and Iteration

  1. 01GPT-6 Operates WeChat for Hours, Showcasing Agent Execution in Long-Chain Tasks
  2. 02Muse Positions Personal Agent as a "Continuous-Action Butler"; Account Access and Trust Are the Hurdles
  3. 03Opus 5.5 Used to Generate Icons and Finished Videos Directly from Prompts; Code-Based Drawing and Iteration is a Reusable Path
  4. 04For Complex AI Video Shots, Do a "Gray-Box Pre-Visualization" First, Separating Space and Camera Movement from the Generation Process
  5. 05Open-Source 4B GUI Operation Model Cua Releases 168 Benchmark Results; Lightweight Approach Worth Watching
  6. 06RADAR Positions Abdominal CT Multimodal Analysis as an Open-Source Model Direction; Training Data and Clinical Positioning Are Focus Points
  7. 07qiaomu home Listed on Obsidian Official Plugin Library, Expanding Note Library into a Daily Work Hub
09.26Sat · No.178
09.25Fri · No.177
09.24Thu · No.176
09.23Wed · No.175
09.22Tue · No.174
09.21Mon · No.173
09.20Sun · No.172

Qwen-Image-2.1 Integrates Image Generation and Editing, Confucius4-R2T2 Optimizes Real-Time Speech Recognition, Anthropic Expands into Wet Labs

  1. 01Qwen-Image-2.1 Combines Open-Source Image Generation, Editing, and Transparent Output in a 7B Model
  2. 02Confucius4-R2T2 Makes Real-Time Speech Recognition More Suitable for Voice Agent Integration
  3. 03Jev Being Explored as a Low-Cost Semantic Judgment Layer for Agents
  4. 04Hypit + Codex Breaks Down Reference Videos into Modifiable Workflows
  5. 05CapCut Integrates AI Generation Capabilities into Its Editor and Creation Agent
  6. 06Reportedly, Anthropic Advances AI R&D into Physical Wet Laboratories
  7. 07Feishu CLI Can Search Public Documents; Personal Cloud Document Permissions Need Re-checking
  8. 08Bilibili's "AI Infinite Arena" Provides Model Evaluation Samples Closer to Real Tasks Than Leaderboards
09.19Sat · No.171

Jev Shifts to Structured Judgment, Alibaba Advances Real-Time Speech Translation, AI Video Tools Move to Editable Workflows

  1. 01Jev Shifts from "Chatting" to "Judging," Entering the Practical Agent Loop
  2. 02Alibaba Advances Real-Time Speech Translation to Low Latency, Multi-Speaker, and Audio-Video Processing Chains
  3. 03OpenClaw 2026.9.5 Bundles Multi-User Collaboration, Browser, and Update Reliability in One Upgrade
  4. 04Claude Code Adopts AGENTS.md, Project Rules Gain Cross-Tool Reusability Path
  5. 05AI Video Tools Shift from One-Time Generation to Modifiable, Engineered Workflows
  6. 06Two Reports Remind: When AI Errors Cross Isolation and Review Boundaries, They Enter Real-World Decision Chains
  7. 07A Blogger's Analysis of Mistral's Strategic Shift: From Frontier Model Competition to Enterprise Deployment and Compute Services
09.18Fri · No.170

Claude Code Projects Enables Multi-Agent Collaboration, Qwen3.8-Omni-Flash Integrates Full-Modal Agent Workflows, OpenAI Astra Advances Legal Workflows

  1. 01Claude Code Projects: Embedding Multi-Agent Collaboration into a Single Project Conversation
  2. 02Qwen3.8-Omni-Flash Integrates Audio-Video Understanding, Reasoning, and Tool Calling into an Agent Workflow
  3. 03OpenAI Advances GPT-6 Astra into Legal Workflows and Industry Plugin Ecosystem
  4. 04Figure Helix 2.5 Advances Humanoid Robot Generalization Testing into Unfamiliar Homes
  5. 05ZCode Source Code Upload Controversy Makes Local Agent Transparency a Product Baseline
  6. 06The "Verifier's Law" Explains Why AI First Conquers Measurable Digital Tasks
  7. 07The Key to Vibe Coding Shifts from Reading Code to Decomposition, Acceptance, and Rollback
  8. 08OpenAI Developer Tools Integrate Context Access and Usage Visualization into a Desktop Experience
09.17Thu · No.169
09.16Wed · No.168
09.15Tue · No.167
09.14Mon · No.166

AI Safety Becomes Industry Consensus, Recursive Self-Improvement Enters R&D Roadmaps, Agent Capabilities Expand from Troubleshooting to Company Creation

  1. 01Frontier AI's "Slowdown" Shifts Toward Pre-Training Safety
  2. 02Zhipu Writes Recursive Self-Improvement into Next-Generation Model Roadmap
  3. 03RSI is Being Deconstructed from a Slogan into a Five-Level Path from Runtime Adaptation to Meta-Improvement
  4. 04Agents Begin to Handle Faults in a Closed Loop at Real Work Sites
  5. 05Grok Bot Galaxy Turns "AI Building Companies" into Public Live Action
  6. 06AI Video Production Shifts from "Card Drawing" to White Model and Reference-Driven Approaches
  7. 07Personal Knowledge Tools Merge Reading, AI Assistance, and Note-Taking into Mobile
  8. 08After Agent Scaling, Bottlenecks Shift to Runtime and Infrastructure
  9. 09The Differentiation Point for AI Application Startups Shifts from Development Speed to Growth and Distribution
09.13Sun · No.165

AI Governance Enters Verifiable Phase, Inference Chip Competition Shifts to Latency Efficiency, Agents Achieve Fault Loop Closure

  1. 01Frontier AI Companies' 'Slowdown' Advocacy Moves into Verifiable Governance
  2. 02Jalapeno Discussion Pulls AI Chip Competition Back to Inference Latency and System Efficiency
  3. 03Agents Evolving from Chat Assistants to Online Fault Loop Closure with Human Gates
  4. 04GPT-6 Astra and GPT Image 2.5 Outputs Shift from Single Images to Multimodal Works
  5. 05Key to AI Visual Workflows Shifts to Scene Assets and Reusable Prompts
  6. 06Independent Developers Are Turning 'Player-Built Worlds' into Sustainable Game Systems
  7. 07LLM Application Cost and Latency Optimization Still Returns to KV Cache Fundamentals
  8. 08As AI Replaces Intermediate Steps, Humans Must Hold Firm on Goal Selection and Outcome Judgment
09.12Sat · No.164

GPT Images 2.5’s Creative Bottleneck Shifts to Consistency, OpenAI Expands Model Capabilities to Research and Voice, Anthropic Proposes Frontier AI Governance Plan

  1. 01GPT Images 2.5: Single-Generation Strength is High, Bottleneck for Continuous Creation Shifts to Consistency
  2. 02OpenAI Rolls Out New Model Capabilities to Research, Voice Interaction, and High-Concurrency Infrastructure
  3. 03The Current Round of Astra/Codex Fixes Targets Reproducible Quality and Attribution Issues
  4. 04Coding Agent Shifts from Q&A Tool to Long-Term Projects and Cloud Execution
  5. 05"Language Input—Agent—Generated Interface" is Becoming a Verifiable Product Direction Judgment
  6. 06A practical monetization path for vibe coding is to first build attention assets, then offer services
  7. 07Anthropic proposes a governance plan to "pace" frontier AI, key lies in external oversight during training
  8. 08AI discovers and validates a WeChat zero-click worm, highlighting the declining cost of cyberattack capabilities
09.11Fri · No.163
09.10Thu · No.162

DeepSeek V4.1 Flash Launches, Agent Operations Become Auditable, Generative Imaging Shifts to Controllable Workflows

  1. 01DeepSeek V4.1 Flash Launches: Official Focus on Vision, Efficiency, and Open Deployment
  2. 02Model and Harness Begin Co-optimization Under the Same Runtime Environment
  3. 03Computer-use Evolves from "Capable of Execution" to Auditable, Interruptible Delivery
  4. 04Generative Imagery Shifts from "Image Output" to Controllable Production Workflows
  5. 05Following iPhone Duo Launch, Development Adaptation and Price Emerge as Practical Barriers
  6. 06AI Safety Discussions Enter Governance Frameworks, But Resignation Statements Remain Personal Accounts
  7. 07Vibe Coding is Shifting Software from Mass-Market Products to Long-Tail Tools
  8. 08Robotaxi Competition Shifts Focus from Demos to Cost, Scale, and Data Loop
  9. 09Paying Gas Directly with USDC: Arc Aims to Reduce Onboarding Friction for New Chains
09.09Wed · No.161

GPT Image 2.5 Full Rollout, GPT-6 Astra Opens, OpenAI Announces Navier–Stokes Solution

  1. 01GPT Image 2.5 Fully Launched, Sketching and Editing Become Direct Entry Points
  2. 02Hands-on Tests Show Image 2.5's Advantages Concentrated in Style Transfer and Controllable Editing
  3. 03GPT-6 Astra Fully Opens, Long Tasks and Visual Operations Become Primary Focus
  4. 04OpenAI Announces Navier–Stokes Result, Mathematical Community Still Requires Independent Verification
  5. 05Discussion Emerges Regarding DeepSeek V4.1 Flash Taking Over V4 Pro
  6. 06Independent Developers Continue Replacing SaaS with Agents, Key Factors Being Cost and Maintainability
  7. 07GEO Conference Places Tools, Tutorials, and Self-Hosted Event Systems in the Same Workflow
  8. 08X Original Content Rewards Opens for Applications, Revenue Sharing Enters Transition Period
09.08Tue · No.160

DeepSeek V4.1 Flash enters beta with price cuts, Xiaomi MiMo Desktop launches work Agent, Astra models shift value to cross-tool delivery

  1. 01DeepSeek V4.1 Flash Enters Internal Beta with API Price Reduction; Speed and Performance Still Require Real-World Testing
  2. 02Xiaomi MiMo Desktop Launches as an Invite-Only Beta for a Work-Oriented Desktop Agent
  3. 03The Value of Astra-Class Models Begins to Manifest as Cross-Tool Delivery, Not Just Single-Round Generation
  4. 04The Key to Agent Engineering Shifts from Prompts to Verifiable Execution Environments
  5. 05The Value of AI Research Agents Lies in Autonomous Exploration, but Goal Definition and Accountability Remain with Humans
  6. 06Obsidian is being transformed into a workbench integrating news reading, RSS, and study notes
  7. 07MiniMax's multimodal workflow begins delivering playable interactive prototypes
  8. 08AIGC visual content attribution disputes drive demand for verifiable creation trails
  9. 09Grok Bot's productization path emphasizes small teams, rapid iteration, and proactive feature removal
09.07Mon · No.159

OpenAI’s Automation Research Milestone, GPT-6/Astra’s Continuous Task Capability Expansion, and Programming AI Covers Formal Mathematics

  1. 01OpenAI Discloses Milestone in Automation Research, But Recursive Self-Improvement Remains Constrained by Safety Boundaries
  2. 02GPT-6/Astra's Computer Use Has Progressed from Single Demonstrations to Continuous Tasks
  3. 03Programming-Capable AI Begins to Cover Formal Mathematics, But Task Orchestration Remains a Bottleneck
  4. 04Codex's New Context Mechanism Shows a Clear Cost-Benefit Trade-off
  5. 05Competitive Focus of Computer-use Tools Shifts to Background Control, Learnability, and Transferability
  6. 06Bilibili's Build in Public Event Demonstrates a User Co-Creation Loop for AI Products
  7. 07GPT-6/Astra's Compute Scale Intensifies Domestic Catch-Up Pressure
  8. 08Anthropic's $1.6 Trillion Valuation is a Model Projection, Not a Confirmed Valuation
09.06Sun · No.158

GPT-6 Astra Achieves Long-Process Desktop Operations, Open-Source Models Focus on Real-Time and Local Deployment, AI Becomes a Life Risk Screening Entry Point

  1. 01GPT-6 Astra's Value Begins to Materialize in "Operating Computers on Behalf of Users"
  2. 02Astra Can Now Execute Long-Form Desktop Workflows, But Deliverable Quality Still Requires Human Oversight
  3. 03The Stronger the Model, the More Critical Harness, Permissions, and State Management Become
  4. 04Model Subscription Value is Beginning to be Measured by "Number of Deliverable Tasks"
  5. 05The Effectiveness of Skills and Agents Increasingly Depends on Model Adaptation and Productization
  6. 06Open Source Models Continue to Advance in Real-Time Interaction and Low-Resource Deployment
  7. 07AI is Becoming the First Screening Entry Point for Specific Life Risks
  8. 08The Boundaries of Originality, Attribution, and Copyright for AI Works Are Still Rapidly Evolving
09.05Sat · No.157
09.04Fri · No.156

GPT-6 Astra Launch Sparks Capability Assessment Debate, Open Source Ecosystem Expands Vertically, AI Safety and Physical World Modeling Take Center Stage

  1. 01GPT-6 Astra Launches: Capability Leap and Evaluation Metrics Must Be Considered Together
  2. 02Astra's Key Changes Lie in Computer Operation, 3D Modeling, and Self-Verifying Workflows
  3. 03WeChat Xiaowei Shows Signals of AI-to-AI Social Internal Testing, Authorization Mechanism is the Core Change
  4. 04Open-Source AI is Evolving from "Releasing Weights" to Platformization and Complete R&D Processes
  5. 05Agent Overreach Incidents and Model Guardrails Turn Safety from Principle to Deployment Condition
  6. 06Vertical Medical Models Begin Transitioning from Doctor Assistants to Foundational Models
  7. 07World Models and Humanoid Robots Simultaneously Push AI Towards Interactivity and the Physical World
  8. 08AI Video Remains an Iterative Engineering Process of Human-Machine Collaboration; Workflow Methods Are More Important Than Single Attempts
09.03Thu · No.155

Frontier Model Competition Enters Hourly Rhythm, Agent Systems Shift to Latent Space Collaboration and Continuous Learning, AI Productization Covers Pre-Market Research

  1. 01Frontier Model Releases Enter High-Frequency, Short-Cycle Competition
  2. 02Fable 5.1's Strengths Lean More Towards Long-Horizon Execution and Scientific Workflows
  3. 03Agents Begin Taking Over Desktops and Business Processes
  4. 04Agent System Competition Shifts Towards Latent Space Collaboration and Continuous Learning
  5. 05Coding Agents Are Evolving from Writing Code to Operating in Scientific and Physical Environments
  6. 06AI is Moving Market Research and Product Decisions Forward
  7. 07Educational Institutions Simultaneously Tighten Usage Boundaries and Rewrite Software Courses
  8. 08The Institutional Boundaries for AI Implementation Are Still Being Rapidly Redrawn
09.02Wed · No.154

Claude Fable 5.1 Significantly Reduces Agent Costs, Qwen3.8-Max-0902 Breaks Programming Benchmarks, Gemini Agentic Video Revolutionizes Long-Form Video Retrieval

  1. 01Claude Fable 5.1 Brings Long-Task Capability, Cost, and Enterprise Privacy to the Forefront
  2. 02Qwen3.8-Max-0902 Refreshes the Agent Programming Competition with Continued Post-Training
  3. 03Gemini API's Agentic Video Shifts Long-Video Retrieval from "Frame-by-Frame Scanning" to Adaptive Forensics
  4. 04World Labs Atlas Advances Video Generation to Controllable 3D World Reconstruction
  5. 05OpenAI Elevates Astra's Cybersecurity Capability to Critical, Simultaneously Lists Monitorability as a Constraint
  6. 06Uber's Software Factory Case Shows Agents Moving from Personal Assistants into Full R&D Pipelines
  7. 07WorkBuddy Open Platform Begins Competing for the Identity and Permission Layer of Cross-Device Agents
  8. 08China's AI Governance Continues Extending Upstream from Content Handling to Models and Agent Applications
09.01Tue · No.153

August 20262026 · 31 issuesAugust 2026 · Visual Briefs →

08.31Mon · No.152
08.30Sun · No.151
08.29Sat · No.150

OpenAI Terminates Cursor Direct Access Partnership, Agent Memory Shifts to Git-based Assets, AI Tools Enter Reviewable Workflows

  1. 01OpenAI to Terminate Cursor's Direct Model Access on November 12, Signaling Shift in Model Supplier Relationships
  2. 02Agent "Memory" is Shifting Towards Git-like Assets and Human-Feedback-Driven Skill Iteration
  3. 03Agent-native Web is Materializing Through MCP, WebMCP, and "Current Interface Context"
  4. 04AI Tools are Moving into Reviewable Professional Workflows, Not Just Chat Interfaces
  5. 05Video Model Research Shifts Focus Towards Low Compute and Real-time Understanding
  6. 06"AI Training AI for Safety" Demonstrates the Potential of Automated Alignment, While Exposing Supervision Boundaries
  7. 07Quantization Enables Massive MoE Models to Enter the ~200 GiB Multi-GPU Deployment Range
  8. 08The Bottleneck for AI-Generated Images and Videos Shifts from "Generation" to Realism, Physical Plausibility, and Controllability
  9. 09Creators Publicly Share Prompts, Skills, and Runnable Works Together, Further Lowering the Barrier to Reuse
08.28Fri · No.149

Domestic Models Shift to Real Workflows, Agents Begin Handling Tasks, Generative Video Barriers Lower

  1. 01Tencent Hunyuan Hy4 Preview Released, Domestic Model Competition Continues to Shift Towards Real Workflows
  2. 02Agent's Focus Shifts from Chat Windows to "Logging In and Handling Tasks for You"
  3. 03Clear Divergence Emerges Between Models That "Can Do" and Those That "Can Run Stably"
  4. 04AI Begins Entering Research Institutions and Lab Equipment, Not Just Researchers' Chat Windows
  5. 05The Barrier to Generative Video Continues to Lower, Shifting Focus to Reusable Production Pipelines
  6. 06Natural Language is Becoming the Primary Entry Point for Game Prototyping
  7. 07Open-Source Robots Begin Forming a "Purchasable—Trainable—Reproducible" Closed Loop
  8. 08AI End-to-End Delivery is Changing How Individuals Create and Develop
  9. 09Agent Architecture Discussions Shift from "More Agents" to Memory, Specialization, and Verifiable State
08.27Thu · No.148

Multi-Agent Security Risks Draw Attention, OpenAI Strategy Returns to Codex and Agent, GLM-5.3-Flash Sparks Cost and Performance Debate

  1. 01OpenAI Publishes Investigation into Hugging Face Incident, Multi-Agent Constraints Emerge as Core Security Issue
  2. 02TIME Interview Summary Indicates OpenAI Refocusing Main Strategy on Codex→Agent
  3. 03GLM-5.3-Flash Becomes Focus of Discussion for Domestic Models, Low Cost and Real-World Experience Still Require Simultaneous Verification
  4. 04Gemini 3.5 Transcribe Advances Real-Time Speech Transcription into a Tool-Callable Workflow
  5. 05Claude Expands into Executable Work Entry Points: Browser Operations and CRM Collaboration Integrated into the Same Ecosystem
  6. 06Grok Bot Begins Offering Cloud Computers with Graphical Interfaces, Further Extending the Agent's Execution Boundary
  7. 07OpenDesign Enters the Design Agent Space with an Open-Source Collaborative Canvas, Shifting Focus from Generation to Deliverability
  8. 08AI Video and Design Tools Enter High-Frequency Creation, but Cost and Deliverability Remain Bottlenecks
08.26Wed · No.147

Domestic Model Opens New Architecture, OpenAI’s Self-Developed Chip Tested, Agents Begin Handling Web Tasks for Users

  1. 01GLM-5.3-Flash Officially Open-Sourced, Domestic Model Pushes Capabilities, Pricing, and Domestic Computing Power to the Forefront Simultaneously
  2. 02Qwen3.8-Flash-Next Open-Sourced as a Qwen4 Architecture Preview, Low Activation Parameters Become a New Direction
  3. 03OpenAI's Self-Developed Jalapeño Inference Chip Enters Practical Testing, Computing Power Competition Shifts to Output Per Watt
  4. 04WebMCP Integrated into ChatGPT Desktop Browser, Websites Begin to Have Tool Interfaces for Agents
  5. 05ChatGPT Work Now Supports Secure Login, Agents Evolve from "Giving Advice" to Completing Web Transactions on Behalf of Users
  6. 06Apodex 1.1 Advances Deep Research into Executable, Resumable, Verifiable Agent Workflows
  7. 07Gear Zero Enables Natural Language to Directly Create Browser Mini-Games, Agents Begin to Handle the Full Production Chain
  8. 08OpenWorker Applies Locally Auditable Agents to Vulnerability, Dependency, and Cloud Configuration Checks
08.25Tue · No.146

Codex Session History Anomaly Draws Attention, OpenAI Restores Plus Usage Limits, Doubao Work Launch Intensifies Office Agent Competition

  1. 01Multiple User Reports Indicate Codex Session History Loss May Be Index or Projection Anomaly
  2. 02OpenAI Announces Reinstatement of 5-Hour Usage Window for Plus, Pro Tiers Temporarily Exempt
  3. 03Tibo Envisions Next-Generation Agents Will Hide Skills, Memory, and Sub-agents
  4. 04Doubao Work Launches, Office Agent Competition Shifts to Organizational Context
  5. 05Agent Search Infrastructure Enters Price War Phase
  6. 06Claude Simultaneously Enhances Long-Output Rendering and Enterprise MCP Permission Infrastructure
  7. 07Chinese Open-Source Models Become Common Base for Alignment Research, Practical Gap is Narrowing
  8. 08NVIDIA and SpaceXAI's Agent Hardware Roadmap Extends to Orbital Computing
  9. 09Keelung District Prosecutors Office Reportedly Indicts Case Involving Illegal Export of 130 B300 Servers
08.24Mon · No.145

AI Programming Tools Reshuffle, Enterprise Procurement Shifts to Cost-Effectiveness, Hugging Face Explores Sale

  1. 01Coding Agents Enter Mainstream Workflows, But the Tool Landscape is Reshuffling
  2. 02DeepSeek Harness Reportedly Has High-Risk Remote Code Execution Vulnerability; Exposure is Key
  3. 03ByteDance Integrates TRAE and Coze into Doubao System, Further Consolidating AI Office Product Line
  4. 04Enterprises Begin Measuring Models by "Good Enough and Cheap"; The Most Powerful Model May Not Drive the Most Spending
  5. 05Hugging Face Reportedly Exploring Sale for Approximately $13 Billion; Value of Open-Source Ecosystem's Distribution Layer Gains Attention
  6. 06Xiaomi's AI Cube Combines Mobile, AI Acceleration, and Autonomous Driving Chips into a Local Compute Host
  7. 07AI Short Video Production is Shifting from Single Works to Reusable Assembly Lines
  8. 08The Difficulty of AI Safety Extends to Boundary Drift in Long Tasks and Offline Dual-Use Capabilities
08.23Sun · No.144

Agent Model Interchangeability Rises, AI Infrastructure Bottlenecks Emerge, Gap Between User Perception and Product Experience

  1. 01Agent Usage Rises, Model Layer Moves Towards Interchangeability
  2. 02Instant Team Joins OpenAI, Agent Backend Infrastructure Becomes a Key Focus
  3. 03Ox Alpha Shows Usable Multimodal and Coding Capabilities in Tests, But Response Efficiency Remains a Weakness
  4. 04Marin 535B-A23B Launches Public Training, The Training Process Itself Becomes a Research Product
  5. 05AI Compute Expansion Simultaneously Hits Power, Water, and Memory Cost Constraints
  6. 06Public Speech Evaluation Exposes Test Set Contamination, Process Data More Valuable Than Leaderboard Scores
  7. 07General Users Are Already Using AI, But Their Mental Model Remains at "Better Chatbot"
  8. 08Image Generation Workflows Shift from Prompt Stacking to Consistency and Visual Directing Systems
08.22Sat · No.143
08.21Fri · No.142

DeepSeek Vision Model Opens API, Anthropic Enhances Enterprise Solutions, GPT-Image-2 Supports Transparent Backgrounds

  1. 01DeepSeek Vision Model Transitions from Code Clues to API Availability, Low-Cost Multimodal Agent Enters Practical Testing Phase
  2. 02Anonymous Model Ox Alpha Shows Strong Coding Capability Signals, But Identity and Benchmark Results Remain Unconfirmed
  3. 03Codex Quota Controversy Not Yet Officially Defined as a Widespread Reduction; Official First Confirms 20 Million Active Users and Issues Resets
  4. 04Anthropic Simultaneously Addresses Enterprise Privacy, Desktop Operations, and AI Education Entry Points, Shifting Product Focus from Capability Demonstration to Adoption Condition Building
  5. 05GPT-Image-2 API Preview Adds Support for Transparent Backgrounds, Making Image Generation More Directly Applicable for Production Assets
  6. 06Anthropic IPO Expectations Heat Up, But Prospectus, Fundraising Scale, and Financial Figures Remain as Unverified Information
  7. 07Xiaomi MiMo-V3-Pro Benchmark Screenshot Allegedly Leaked Early, Domestic Agent Model Competition Continues to Approach the Flagship Segment but Awaits Official Announcement
  8. 08The Value of Travel Agents Begins to Shift from "Planning for You" to "Executing Operations for You," But Effectiveness Depends on Industry System Integration
08.20Thu · No.141

AI-designed cancer vaccine shows breakthrough, new robot learning paradigm emerges, validation framework boosts model cost-effectiveness

  1. 01AI-Designed Personalized mRNA Cancer Vaccine Shows Positive Signal in Melanoma Phase III Trial
  2. 02OpenAI Previews Private Safety Processing, Continues Zero Data Retention for Frontier Models
  3. 03Multimodal Capabilities of Coding Agents Begin Integrating into Context and Tool Loops
  4. 04Agent Skills Shift from "Recommendation" to Quantified Evaluation and Safety Checks
  5. 05A Single Demonstration Enables Robots to Combine Actions, Bringing Physical Prompting into the Real World
  6. 06Validation Frameworks Are Turning the Performance and Price Advantages of Low-Cost Models into Tangible Results
  7. 07As AI Enters Education, Tension Emerges Between Product Expansion and Learning Outcomes
  8. 08Local Backlash Against AI Data Centers Enters US Election Discourse
08.19Wed · No.140

OpenAI Pauses Training Due to Rapid Capability Growth, Claude Protein Design Enters Experimental Validation, Agent Products Shift to Cross-Platform Execution

  1. 01OpenAI Temporarily Halts Frontier Reinforcement Learning Training Due to Rapid Capability Growth
  2. 02Claude's Protein Design Progresses from Model Output to Wet Lab Validation
  3. 03DeepSeek Web Interface Shows Signs of Possible New Model Gray Release, Awaiting Official Confirmation
  4. 04GLM-5.3 Competition Shifts Focus to Cost Per Task
  5. 05Agent Products Are Evolving from Answering Questions to Cross-Platform Task Execution
  6. 06H200 Begins Entering Chinese Supply Chain, Introducing New Variables to Training Compute Constraints
  7. 07AI Observatory Brings Real-World Usage Back into Model Evaluation Discussions
  8. 08AI Competitive Barriers Shift from Model Access to Data, Processes, and Organizational Execution
08.18Tue · No.139

DeepSeek V4 Performance Questioned, WeChat Work Opens Office Capabilities to AI, Low-Cost Models Shift to Multi-Generation Self-Verification

  1. 01J-Space's DeepSeek V4 Performance Claims Face Reproducibility Questions
  2. 02WeCom Begins Opening Office Capabilities to External AI Agents
  3. 03Improvement Path for Low-Cost Models Shifts to "Multiple Generations + Self-Verification"
  4. 04Coding Agents Are Evolving from "Writing Code" to Continuous Execution and Editable Design
  5. 05Remotely Controlling PCs and Cloud Agents via Phone Is Becoming a Reusable Work Pattern
  6. 06Seedance 2.5's Video Production Focus Shifts to Pre-visualization, Segmented Generation, and Editing Transitions
  7. 07The Core of AI Outfitting Shifts from "Looking Good" to Simultaneously Locking Character and Garment Identity
  8. 08Agent Products Begin Packaging "Role, Memory, and Tools" into Long-Term Reusable Work Units
08.17Mon · No.138

Stripe acquisition of OpenRouter rumors draw attention, DeepSeek-V4 price hike impacts developer decisions, AI watermark removal tool gains popularity

  1. 01Reports of Stripe's Acquisition of OpenRouter Circulated by Multiple Bloggers, Highlighting the Infrastructure Value of the Model Routing Layer
  2. 02DeepSeek-V4 Price Increase Begins to Impact Token Plans and Developer Usage Decisions
  3. 03Codex Now Offers ChatGPT Users Access to GPT-5.6 Sol's Million-Token Context Configuration
  4. 04OpenAI Demonstrates How "Smarter Model Selection" Can Reduce Agent Costs
  5. 05Multi-Agent Long-Term Memory Introduces "Instruction Propagation" Security Aspect; Research Suggests Adding System-Level Warnings
  6. 06AI Watermark Removal Tools Gain Popularity, Sparking Tension Between Content Provenance and Evasion
  7. 07Vercel Integrates GLM 5.3 into AI Gateway, Bringing Cost-Performance Comparisons of Open-Source Models into Platform Procurement
  8. 08OpenLess Recommended as Open-Source Voice Input Alternative, Focusing on Organizing Dictation into Usable Prompts
08.16Sun · No.137

AI Autonomous Research Demonstrates Experimental Execution Capability, DeepSeek Harness Builds Tool Loop, Video Generation Enters Full Narrative Stage

  1. 01The Key Progress in Autonomous AI Research Lies in "Conducting Experiments," Not Just Generating Ideas
  2. 02The Early Value of DeepSeek Harness Lies in Forming a Closed Loop Connecting Boundaries, Tools, and the Plugin Ecosystem
  3. 03Qwen3.8 is Simultaneously Advancing Local Inference Efficiency, Low-Activation Architecture, and Usable Products
  4. 04The Core of Claude's Text Watermark is Altering the Sampling Random Source, Not Inserting Hidden Characters
  5. 05Codex Multi Agents v2 Delegates Model Selection to Task Routing
  6. 06Seedance 2.5's Capability Enhancement Has Reached "Can Produce Complete Videos," with Price as the Main Constraint
  7. 07Visible Cases of Grok 4.6 and Grok Build Concentrate on "From Chat to Execution"
  8. 08dots3-note preview Brings Long-Horizon Agent Training into Open-Source Model Discussions
08.15Sat · No.136

DeepSeek Harness Shifts to Self-Improving Runtime, OpenAI Growth Focus Moves to Enterprise Workflows, Agent Production Deployment Faces Permission and Tracking Challenges

  1. 01DeepSeek Harness's Focus Shifts from "Usability" to a Self-Improving Runtime and Ecosystem
  2. 02The "Multiple Versions" Narrative for DeepSeek V4 Pro Remains at the Level of Community Observation and Unverified Hypotheses
  3. 03OpenAI's Growth Narrative is Shifting from Consumer Subscriptions to Enterprise Workflows
  4. 04After Agents Enter Production, Key Issues Become Permission Boundaries, Failure Tracking, and Decision Auditing
  5. 05ChatGPT is integrating external materials and computer activity into a continuous context
  6. 06After X's recommendation algorithm was open-sourced, discussion focus shifts to ranking mechanisms and auditability
  7. 07Competition in AI video generation is beginning to manifest in deliverable clips, asset accumulation, and unit cost
  8. 08Generative text watermarking emphasizes "detecting generation traces," but strength depends on text length and rewriting
08.14Fri · No.135
08.13Thu · No.134

DeepSeek V4 Pro Relaunched with Major Price Adjustments, Harness Open-Source Preview Reconstructs Agent Runtime, Claude Browser Agent Achieves Cross-Device Synchronization

  1. 01DeepSeek V4 Pro 0813 Re-released After Rollback, Agent Capability Upgrades and Significant Price Adjustments Emerge Simultaneously
  2. 02DeepSeek Harness v0.1 Open Source Preview, Deconstructs Agent Runtime into Composable Plugins
  3. 03Harness Quickly Enters Practical Testing Phase, Transparent Traces and Cache Costs Become Early Highlights
  4. 04Discussion on Grok 4.6 Entering the Top Tier, Advantages Concentrated in Speed, Cost, and Agent Programming
  5. 05Claude Browser Agent Begins Cross-Device Session Sharing, While Bringing Web Page Injection Risks to the Forefront
  6. 06Codex and Enterprise Agent Usage Continue to Expand, with Linux, Skills, and Plugins Becoming Key Entry Points
  7. 07Effective Paths for AI Video Generation Remain Short Shots, Reference Images, and Revisable Workflows
  8. 08Cordis Paper Grounds the Challenge of Self-Evolving Agents in Revocable Components and Local Recovery
08.12Wed · No.133

DeepSeek V4 Pro Launches, xAI Advances Models and Executable Agents Simultaneously, AI Video Competition Shifts to Continuous Systems

  1. 01Signs of DeepSeek V4 Pro 0813 Launch Emerge, with Agent Capabilities and Pricing as Key Focus
  2. 02xAI Advances Grok 4.6 and Grok Bot Simultaneously, with Model Upgrade and Executable Agent Both Rolling Out
  3. 03ChatGPT and Codex Desktop Preview Expands to Linux, with Workflow Migration and Sync Support
  4. 04A Paper Raises Security Alerts Regarding Cross-Model Compatibility of Encrypted Chain-of-Thought
  5. 05AI Video Evaluation Criteria Shift from Single-Shot Image Quality to Continuous Characters, Dynamic Systems, and Production Costs
  6. 06Agent Work is Shifting from "Capable of Generation" to Being Testable, Orchestratable, and Reusable
  7. 07Public Signals Emerge of Manus Spinning Off from Meta and Resuming Independent Operations
08.11Tue · No.132

Closed-Source Model Chain-of-Thought Protection Bypassed, Claude Advances Riemann Zeta Function Lower Bound, AI-Designed Phage Validated in Lab

  1. 01Researchers Bypass Protection Mechanisms of Closed-Source Models That Hide Chain-of-Thought
  2. 02Claude Advances Known Lower Bound on Riemann Zeta Function Related Problem from 41.6% to 67.2%
  3. 03OpenAI Delivers High-Risk Cybersecurity Capabilities to Defenders in a Controlled Manner
  4. 04Claude Begins Adding Machine-Detectable Source Markers to Text and Files
  5. 05AI-Designed Bacteriophages Validated in Lab for Replication and Infection
  6. 06Frontier AI Risk Discussion Reaches the Level of U.S. Congressional Open Letters
  7. 07Electricity and GPUs Are Evolving from Computing Equipment into Financeable, Revaluable Infrastructure Assets
  8. 08Chinese Large Model Commercialization Accelerates, But Revenue Deposition Still Depends on Inference Capacity and Software Ecosystem
08.10Mon · No.131
08.09Sun · No.130

Seedance 2.5 Shifts to Controllable Video Production, AI Lowers Small Software Delivery Threshold, Agent Competition Expands Focus

  1. 01Seedance 2.5 is Evolving from "Effect Showcase" to a Controllable Video Production Pipeline
  2. 02Viko Launches, Linking Visual Inspiration Deconstruction and Creative Asset Accumulation into a Single Chain
  3. 03AI is Lowering the Delivery Barrier for Small-Scale Software and Product Iteration
  4. 04Agent Tool Competition is Extending into Conversational Collaboration, Plugin Interoperability, and Quota Management
  5. 05The Expansion of AI Capabilities is Also Raising Security Review Requirements for Software, Agents, and Biological Research
  6. 06The Bottleneck for Large-Scale AI Adoption is Emerging as a Physical Supply Chain and Systems Engineering Problem
  7. 07Discussions on X About AI-Generated Junk Content and Automated Spam Reflect That Information Noise is Becoming a Platform Problem
  8. 08The Value of World Models and AI Assistants is Returning to Real-World Tasks and Human Decision-Making Responsibility
08.08Sat · No.129

OpenAI designates Astra as key security model, acquires NextSlide for AI-generated presentations, X platform adjusts original content reward rules

  1. 01OpenAI Lists Astra as First Cybersecurity "Critical" Model; Release Cadence to Prioritize Safety Controls
  2. 02Multiple Bloggers Report OpenAI's Acquisition of NextSlide, Making AI-Generated Editable PPTs a Visible Product Direction
  3. 03X's Original Content Rewards Program Rule Changes Take Effect, Shifting Earnings Basis from Reach to Original Contribution
  4. 04Seedance 2.5 Testing Shows Video Generation Closer to a Production Tool, but Character Consistency and Costs Still Require Human Oversight
  5. 05Claude Code Adds Cross-Session Messaging, Allowing Multiple Tasks to Exchange Summaries and Collaborate in Work Trees
  6. 06AI Acceptance and Testing May Precede Production Automation, with R&D Roles Shifting to Defining Goals and Building Reusable Processes
  7. 07Cloudflare Reportedly Launches Kitesurf Browser for Agents, as Browsers Reduce Interface Burden for Humans
  8. 08Experimental Method Emerges for Running Kimi K3 on 8GB RAM, but Speed and Storage Costs Keep it a Technical Toy
08.07Fri · No.128

OpenAI Launches Reasoning Intensity Adjustment and Cross-Client Plugin Standard, Haidian Launches AI Innovation Belt Urban Design Solicitation

  1. 01OpenAI Turns GPT-5.6 Capability Differences into Direct Product Toggles
  2. 02Agent Plugins Aim to Turn Skills and MCP into Cross-Client Reusable Plugins
  3. 03Codex Security Review Incorporates Repository Context into Pull Request Security Checks
  4. 04Beijing Haidian's AI Innovation Belt Described as an Agent-Only Real Urban Design Solicitation
  5. 05Seedance 2.5's Competitive Focus Shifts to Long Video, Continuity, and Workflow Entry Points
  6. 06Cloudflare is Transforming Website Access and Browser Runtime Environments into Agent Infrastructure
  7. 07Anthropic Relaxes Fable 5's Biosafety Interceptions, But Maintains Dual-Use Research Boundaries
  8. 08Recap of Hugging Face Incident from Multiple Accounts Highlights Collaborative Security Risks of Agent Clusters
  9. 09Multi-Agent CAD Breaks Down Printable Model Generation into Planning, Design, Coding, and Quality Control
  10. 10Alibaba Cloud Releases Qwen 3.8-Max, Continuing to Push Large Model Competition Towards Ultra-Large Scale and Long-Range Tasks
08.06Thu · No.127

DeepSeek API Price Hike Triggers Cost Reevaluation, Wan3.0 Beta Expands Video Generation Context, Google Veterans Found Automated Research Company

  1. 01DeepSeek API to Increase Prices Overall, Low-Cost Benefits Enter Reevaluation Period
  2. 02Wan3.0 Enters Public Beta, Video Generation Begins to Incorporate Document and Web Context
  3. 03SeedRealtime Advances Real-Time Video Calls to Full-Modal, Full-Duplex Interaction
  4. 04Google DeepMind Adjusts Management Roles, Hassabis Shifts Focus to Long-Term Science and AGI
  5. 05Google Veterans Jeff Dean and Others Found Discovery Loop, Targeting Automated Scientific Research
  6. 06Meta Releases Muse Code, Coding Agent's Price Advantage Comes with Data Usage Cost
  7. 07Prime Agent Treats the Agent Framework Itself as a Performance Variable, Results Still Require Independent Review
  8. 08Video Model Testing Shows Selection Should Be Based on Shot Task Specialization, Not Just Overall Rankings
08.05Wed · No.126

AI Model Benchmarks Become Key for Selection, Agent Infrastructure and Design Collaboration Advance

  1. 01Benchmarking Seedance 2.5 vs. MiniMax H3: Cost and Quality Comparisons Now Possible with Identical Prompts
  2. 02Cloudflare Turns Agent "Money" and Email Operations into Callable Infrastructure
  3. 03Open Design Enters Codex Official Plugin Directory, Agents Begin Collaborating on Real-Time Canvases
  4. 04Qwen 3.8-Max Brings Capability Release, Open-Source Preview, and Price Competition Simultaneously
  5. 05News on DeepSeek, Kimi, and SSI Funding and New Models Should Still Distinguish Between Reports and Confirmations
  6. 06Skills Shift from "The More, the Better" to Manual Curation, Testing, and Monetizable Knowledge
  7. 07Model Engineering Begins Discussing Structured Input, Context Management, and Data Quality on the Same Level
  8. 08Agent Applications Show More Specific Internal Use and Real-Environment Test Cases
08.04Tue · No.125

OpenAI Unveils GPT-Live Voice Architecture, Discloses Internal Model Math Research; DeepSeek V4 Flash Low-Cost Strategy Gains Attention

  1. 01OpenAI Unveils GPT-Live Voice Architecture, Focusing on Low Latency and Asynchronous Reasoning
  2. 02OpenAI Discloses Internal Model's Mathematical Research Results and Opens Formal Materials for Verification
  3. 03Cloudflare Developments Point to Agent Execution Environment, Payment, and Identity Infrastructure
  4. 04DeepSeek V4 Flash's Low Price and Short-Term Free Activity Shift Competition Focus Towards Adoption
  5. 05Seedance 2.5 Extends from Content Generation to Industrial Data Synthesis, But Cost Remains a Barrier
  6. 06WorkBuddy Discussion Shifts from Single Tool to Internal Enterprise Agentization and Ecosystem Roles
  7. 07Agent Engineering Practices Begin Emphasizing Planning, Execution, and Verification Loop
  8. 08Agents Capable of Operating External Applications Should Leave Transactions and High-Risk Actions to Humans
08.03Mon · No.124

Qwen3.8-Max Release & Testing, MiniMax H3 Open-Source Restrictions Draw Attention, Claude Demonstrates New Team Collaboration Form

  1. 01Qwen3.8-Max Officially Released, Flagship and 27B Small Model Weights to be Open-Sourced
  2. 02Qwen3.8-Max Long-Task Testing Shows: Capable of Cross-File Iteration, But Still Requires Human Review
  3. 03After MiniMax H3 Open-Source, Running It, Running It Fast, and Using It Legally Are Three Separate Things
  4. 04Seedance 2.5's Official Manual Advances Video Generation into a Reusable Workflow
  5. 05AI Fashion Videos Are Being Validated as a Low-Cost E-commerce Content Production Method
  6. 06Claude's New Form Resembles a Long-Term Collaborator Within a Team, Not Just a Chat Window
  7. 07Individual Developers Begin Solidifying Vibe Coding into Their Own Business Systems
  8. 08Model Usage Strategy Shifts from "Strongest Throughout" to Task Division and Cost Routing
08.02Sun · No.123
08.01Sat · No.122

OpenAI Astra Advances Mathematical Research, DeepSeek V4 Flash Goes Open-Source, YC Open-Sources Organizational-Grade Agent Framework

  1. 01Public Clues for OpenAI Astra Point to "AI-Led Mathematical Research", But Progress Should Still Be Considered Unverified
  2. 02DeepSeek V4 Flash 0731 Pushes High-Capacity Models Further Toward Open-Source, Low-Cost, Local Deployment
  3. 03YC Open-Sources QM: Agent Competition Shifts From Personal Assistants to Organizational Work Infrastructure
  4. 04Codex Plugins Compress the "Manual Glue" in SaaS Launch Workflows Into a Single Conversation
  5. 05Seedance 2.5 Improvements Focus on Realism and Physics, But Creators Still Choose 2.0 Based on Cost
  6. 06Grok Imagine 1.5 Adds Text-to-Video, Image and Audio Reference, Locks 1080p Output Behind Higher Subscription Tiers
  7. 07Google Earth's Nano Banana Experiment Exposes Content Safety Boundaries for Real Location Generation

July 20262026 · 30 issuesJuly 2026 · Visual Briefs →

07.31Fri · No.121

DeepSeek V4-Flash Public Beta Enhances Agent Programming, OpenAI Drives Cost Competition with Major Price Cuts

  1. 01DeepSeek V4-Flash Official API Public Beta: Low-Cost Agent Programming Emerges as the Day's Most Definitive Product Change
  2. 02OpenAI Slashes GPT-5.6 API Prices, Auto-Review Costs Drop in Sync
  3. 03Seedance 2.5 Launch Brings Longer, More Controllable Video Generation, But Actual Cost Proves a Major Hurdle
  4. 04Codex Image Agent Adds Dedicated Editing Interface, Bringing Batch Modifications and Local Retouching Closer to Workflows
  5. 05Anthropic Discloses Three Unauthorized Access Incidents in Evaluation Environment, Shifting Agent Security Discussion from Abstract Risk to Visible Cases
  6. 06OpenAI Engineer Recruitment Features Agentic Coding Round, AI Collaboration Skills Begin Entering Engineering Assessments
  7. 07FLUX 3 Preview Opens Trial via Hermes Agent, Image Models Begin Direct Entry into Short Film Workflows
  8. 08OpenClaw Releases Monthly Extended-Stable Version, Stability and Observability Maturity Take Center Stage in Product Narrative
07.30Thu · No.120

Long-Range Agent Evaluation Reveals Harness and Model Are Equally Important, OpenAI Opens Advanced Models to Researchers, AI Video Models Shift to Reusable Content Production

  1. 01GPT-5.6 Sol Performance Reiterates: Long-Range Agent Evaluation Tests Both Model and Harness
  2. 02OpenAI Opens Advanced Models to Researchers, Extending Product Entry into Scientific Workflows
  3. 03Increased Signals of 'Work' Product Competition, But Team Integration Claims Remain Social Media Information
  4. 04AI Video Models Shift from Version Announcements to Reusable Content Production Templates
  5. 05AI Assistants Face 'Visual Fraud' Risk: The Webpage Humans See May Differ from What the Model Reads
  6. 06ChatGPT/Codex Login Verification Shows Practical Clues for Switching to Authenticator, Not Yet a Universal Update
  7. 07Pitfall Records for eSIM and 1500G Pocket WiFi, Core Risks in Activation, Roaming, and Refund Terms
  8. 08'Doorstep Humanoid Robot' Cases May Still Rely on Human Remote Control, Smooth Actions Don't Equal Autonomy
  9. 09AI Confidence Cannot Replace Human Boundary Judgment, Convenience and Responsibility Awareness Rise Together
07.29Wed · No.119
07.28Tue · No.118

Anthropic Pirated Training Data Ruling, XPeng AI Head Joins OpenAI, NVIDIA Executive Detained in Taiwan

  1. 01Anthropic Training Data Piracy: Court Rules Company Downloaded Over 7 Million Books Despite Knowing Illegal Source
  2. 02XPeng AI Infrastructure Head Joins OpenAI to Lead Embodied Intelligence Work
  3. 03NVIDIA Executive Detained by Taiwan Prosecutors: AI Server Smuggling Case Expands
  4. 04Cursor Accused of Silently Uploading User Source Code by Default: 63,106 Files Transferred Without Explicit Consent
  5. 05Kimi K3 Fully Open Sourced and Integrated with Alibaba Zhenwu PPU: Deep Integration Between Chinese AI Models and Domestic Chips
  6. 06Grok 4.6/4.7 Roadmap Announced: 1.5T and 2.1T Parameter Models to Launch Sequentially
  7. 07Open Secure AI Alliance Launches, Anthropic Makes Restrictive Statement on Open Source Ecosystem
  8. 08SSI and NVIDIA Announce Strategic Partnership: Compute Power to Increase 10x in 12 Months
07.27Mon · No.117
07.26Sun · No.116

ChatGPT Work Active Users Surpass Codex, AI Open Weights Statement Released, Andrew Ng Open-Sources Local AI Framework

  1. 01OpenAI Enters the "Singularity Era," ChatGPT Work Active Users Surpass Codex
  2. 02Jensen Huang's First X Post: NVIDIA Signs Open Weights Declaration, Anthropic Stands as Silicon Valley's Sole Absentee
  3. 03Andrew Ng Open-Sources OpenWorker: A Local-First AI Colleague Framework
  4. 04Midjourney Acquires Astrology App Co-Star, Forms Consumer Product Team
  5. 05Reuters: OpenAI Internal Agent Left "Jailbreak Notes" for Subsequent Instances
  6. 06SpaceX Starship Completes 13th Flight Test
  7. 07Grok Build Continues to Gain Momentum: Voice Interaction and Grok Imagine Feature Released
  8. 08Claude Suspends New User Registrations Due to Abuse; Azure China Region Launches Opus 5
  9. 09DeepSeek V4-Pro Remains the Most Cost-Effective Agent Model
07.25Sat · No.115
07.24Fri · No.114

Chinese Mathematicians Win First Fields Medal, Huang Renxun and Tech Giants Back Open-Source AI, FLUX 3 Unifies Multimodal Generation

  1. 01Wang Hong and Deng Yu Win Fields Medal: Chinese Mathematicians Secure Mathematics' Highest Honor for the First Time
  2. 02Jensen Huang Joins X, Co-signs Open Letter with Tech Giants to Support Open-Source AI Ecosystem
  3. 03FLUX 3 Officially Released: A World Model Unifying Image, Video, Audio, and Motion Prediction
  4. 04Cognition Acquires Poke: Devin Enters the Era of Always-On General-Purpose Agents
  5. 05Anthropic Invests Another $20 Million in AI Regulation Lobbying, Escalating Tech Giant Political Games
  6. 06OpenAI Launches GPT-Live Voice Desktop Control and Multiple Codex Updates
  7. 07AI Chip Company Etched Completes $300 Million Series C Funding, Valuation Exceeds $10 Billion
  8. 08Claude Opus 5 Showcases Impressive 3D Capabilities, Multiple AI Security Incidents Continue to Unfold
07.23Thu · No.113

DeepSeek’s Philosophy of Restraint Sparks Debate, US Sanction Threats Intensify, Chinese Open-Source Models Reach 58% Share

  1. 01DeepSeek Founder's 'Philosophy of Restraint' Goes Viral After Leaked Investor Meeting Content is Deleted
  2. 02US Threatens Sanctions on Chinese AI Firms; White House AI Advisor and OpenAI Publicly Clash
  3. 03Data Shows Chinese Open-Source Models Account for 58% of AI Token Usage by US Companies
  4. 04GPT-5.6 and Grok 4.5 Solve Two ~30-Year-Old Mathematical Conjectures on the Same Day
  5. 05Kimi K3 Matches Fable 5 in Agent Tasks, Model Routing Can Cut Costs by Up to 50x
  6. 06Codex to Get Real-Time Voice Mode, Opus 5 Confirmed, Flux 3 Released Open-Source
  7. 07Human Go Player Shin Jin-seo Defeats AI KataGo in Handicap Match Comeback
  8. 08Amazon Cuts AGI Department; US Government Launches $5 Billion "Genesis Plan"
  9. 09GPT-5.6 Exhibits Reward Hacking; Agent Programming Mired in "Shell" Marketing Controversy
07.22Wed · No.112
07.21Tue · No.111

AI Achieves Perfect IMO Score, Kimi K3 Enters Azure, US Considers Ban

  1. 01AI Fully Conquers IMO 2026, Multiple Cutting-Edge Models Achieve Perfect Scores
  2. 02Kimi K3 Enters Microsoft Azure Supply Chain, Chinese Models Move into Global Infrastructure
  3. 03US Government Considers Ban on Chinese Open-Source Models, K3 Release Intensifies Policy Debate
  4. 04zAI Completes 1GW Domestic Chip AI Data Center, Compute No Longer a Bottleneck for Chinese Models
  5. 05Google Advances Dedicated Inference Chip Frozen v2, AI Chip Competition Enters New Phase
  6. 06Jensen Huang Publicly Admits US Chip Ban Failure, Huawei's Rise as Key Example
  7. 07OpenAI Internal Model Access Suspended Due to Alignment Issues, Highlighting Agent Era Security Challenges
  8. 08Unitree Releases UniFoLM-OmniA-0.3 Embodied Model, Robots Evolve Towards True AI Agents
  9. 09Huawei Kirin 9030 Chip Teardown: Metal Pitch Surpasses Intel 18A, No EUV Lithography Used
  10. 10Thinking Machines Lab Continues Expanding Pre-training Team
  11. 11Viral 'Chinese LLM Distillation Controversy' PDF Confirmed as Written by Middle Schooler Using QQbot
07.20Mon · No.110

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

  1. 01Claude Fable 5 Disproves Jacobian Conjecture, AI Cracks 87-Year Math Problem During World Cup
  2. 02Hugging Face Breached by Fully Automated AI Agent; Closed-Source Models Refused to Assist in Investigation, Open-Source GLM 5.2 Completed Analysis
  3. 03Kimi K3 and the US-China AI Gap Discussion; Qwen3.8-Max and Grok 4.5 Make Moves
  4. 04OpenChatCut Released: An Agent-native, Local-first Open-Source AI Video Editor
  5. 05Pinduoduo Recommendation Algorithm Sparks Privacy Debate; Follow-up Controlled Experiment Points to Algorithm Bug
  6. 06WeChat Channels Download Tool and learnui Website Resource
  7. 07New Progress in Anti-Aging Research: Blocking EP2 Receptor Restores Immune Clearance Ability
  8. 08Model Fingerprinting, Codex Reset Controversy, and Other Practical Info
07.19Sun · No.109
07.18Sat · No.108
07.17Fri · No.107
07.16Thu · No.106

Kimi K3 Open-Source Model Released with Performance Nearing Top-Tier; Grok Build and Coding Tools Successively Open-Sourced or Reset Quotas

  1. 01Kimi K3 Officially Released: Moonshot AI Launches 2.8T Parameter Open-Source Model, Multiple Benchmarks Nearing International Top-Tier
  2. 02Grok Build Officially Open-Sourced, xAI Resets All User Quotas
  3. 03Coding Tools Compete to Reset Quotas: Claude and Codex Lift Usage Limits on the Same Day
  4. 04Anthropic Partners with Top Investors to Launch FDE Company Ode, Model Makers Boost Enterprise Deployment Services
  5. 05Thinking Machines Releases Inkling: Mira Murati Team's First Open-Source Multimodal Large Model
  6. 06OpenAI Launches First Hardware Product: kbd-1.0-codex-micro Custom Keyboard, Priced at $230
  7. 07OpenAI Releases GPT-Red: Internal Automated Red Team System Enhances Model's Resistance to Prompt Injection
  8. 08Anthropic's National Security Policy Director Comments on US-China AI Competition: GLM 5.2 is Currently China's Most Advanced Model
07.15Wed · No.105
07.14Tue · No.104
07.13Mon · No.103
07.12Sun · No.102

GPT-5.6 Sol Achieves Top Rankings Across Benchmarks, Noise Image Tests Expose Model Variances, AI Video Generation Advances to Precision Control

  1. 01Noise Image Probing AI Model 'Mind Reading': Significant Differences in Model Reactions to Nonsensical Images
  2. 02GPT-5.6 Sol Tops Comprehensive Evaluations, Ultra Layer Slowed by Superpowers Plugin
  3. 03Grok 4.5 Heavily Promoted by Musk, Video Generation and Low-Cost Subscription Draw Attention
  4. 04GLM 5.2 Shows Best Cost-Performance in Databricks Internal Evaluation
  5. 05Starship 13th Test Flight Targets July 16, First V3 Starlink Deployment
  6. 06ChatCut Codex Plugin Virality Sparks AI Product Open Ecosystem Debate
  7. 07AI Video Generation Enters Era of Fine Control: Gemini Omni and Grok Showcase Strengths
  8. 08Agent Infrastructure Evolution: Orchestration Layer Thinning, ROI Measured from Individual Level
07.11Sat · No.101
07.10Fri · No.100
07.09Thu · No.99

Grok 4.5 Challenges Programming Benchmarks, OpenAI Dense Updates on Voice and Models, ByteDance Launches Cost-Effective Image Model

  1. 01SpaceXAI Officially Releases Grok 4.5, Leading in Multiple Benchmarks
  2. 02OpenAI Releases GPT-Live and Previews GPT-5.6 on the Same Day, Reports Indicate Accelerated GPT-6 Development
  3. 03OpenAI Withdraws Recommendation After Auditing SWE-Bench Pro, Sparking Industry Discussion
  4. 04ByteDance Releases Seedream 5.0 Pro Image Model, Multiple Evaluations Conducted
  5. 05China AI Industry Updates: Claude Code Risk Warning, Baichuan Nia Teams, Stepfun Agent Phone, Xiaohongshu REDSkill
  6. 06MiniMax Launches Token Plan Subscription Package, M3 Pro Flagship Model Incoming
  7. 07OpenClaw Foundation Announces Establishment, Grok 4.5 Launches Simultaneously
  8. 08Overview of Other AI Industry Developments
07.08Wed · No.98
07.07Tue · No.97

Anthropic discovers model’s internal ‘thinking space’, DeepSeek develops own inference chip, China plans to restrict overseas access to AI models

  1. 01Anthropic Discovers Model's Internal 'Thinking Space' J-space, Sparking AI Consciousness Debate
  2. 02DeepSeek Reportedly Developing Its Own Inference Chip, Strategic Shift to Hardware
  3. 03China Plans to Restrict Overseas Access to Domestic Top AI Models
  4. 04Li Bojie, 'Huawei Prodigy' Founder, Exposes Investor Dispute Involving Metagent Project
  5. 05Claude Code Product History Released: From Internal Tool to $500M Annual Revenue AI Coding Platform
  6. 06motion-anything Open-Source Release: Free Open-Source Alternative to Figma Motion
  7. 07X Platform Updates: Renamed to SpaceXAI, Bot Attack Giving Memberships, Growth Data Hits New High
  8. 08Fable Frontend Capability Showcase: Generate 3D Interactive Page with One Sentence
07.05Sun · No.96
07.04Sat · No.95
07.03Fri · No.94
07.02Thu · No.93
07.01Wed · No.92

June 20262026 · 30 issuesJune 2026 · Visual Briefs →

06.30Tue · No.91
06.29Mon · No.90
06.28Sun · No.89
06.27Sat · No.88
06.26Fri · No.87
06.25Thu · No.86

Anthropic Alleges Largest-Ever Distillation Attack by Alibaba, ByteDance Releases Seed-2.1-Pro, Qualcomm Acquires Modular

  1. 01Anthropic Publicly Accuses Alibaba's Qwen of Launching Largest-Ever Distillation Attack
  2. 02Apple Raises Prices Across Entire Product Line, Covering Mac and iPad Series
  3. 03Claude Fable 5 Shows Signs of Return, But Actual Restoration Awaits Confirmation
  4. 04ByteDance Releases Doubao-Seed-2.1-Pro, Multiple Developers Test Coding Capabilities
  5. 05OpenAI Releases Custom Chip Jalapeño and GPT-5.5 Instant Update
  6. 06Qualcomm Acquires AI Computing Firm Modular for Approximately $3.92 Billion in All-Stock Deal
  7. 07Samsung Electronics Officially Ends Home Appliance Retail in China, Official WeChat Account Deregistered
  8. 08Google Gemini Loses More Core Members, Two Join Anthropic
  9. 09Tencent Launches Agent-Specific Email, Feishu CLI + Codex Workflow Emerges
06.24Wed · No.85
06.23Tue · No.84
06.22Mon · No.83

WeChat Agent ‘XiaoWei’ Opens Beta, AI Browser Plugin Trojan Affects 900K Users, Anthropic Completes New Mythos Training

  1. 01WeChat Agent 'XiaoWei' Begins Beta Testing, Capabilities Far Exceed Expectations
  2. 02AI Browser Plugin Trojan Incident: Attackers Buy Search Ads, Over 900K Users Affected
  3. 03Anthropic Completes Training of New Mythos Model, Whether to Offer Externally Not Yet Determined
  4. 04Alibaba Management Team-Building in Hangzhou Signals Stability; Happy Horse 1.1 Video Model Released
  5. 05Sakana AI Releases Fugu Model Orchestration Scheme: Single API Schedules Multiple Expert Models
  6. 06Codex / Claude Code Ecosystem Continues to Expand: Token Tracking, Video Editing, Multi-Round Review and Other Practical Tools Emerge
  7. 07GPT-Image-2 Breakthrough in Generating Architectural Analysis Diagrams: From Drawing to Understanding Spatial Logic
06.21Sun · No.82

DeepSeek Harness Group Launches Large-Scale Hiring, Mythos Breaches Pentagon Systems, SubQ 1.1 Small Enhances Attention Efficiency

  1. 01DeepSeek Harness Group Launches Large-Scale Hiring, Multiple Positions Open
  2. 02Codex Infinite Canvas Image Generation Plugin Cowart Open-Sourced, Community Quickly Follows
  3. 03Doubao Updates AI Outfit Video Capability, Jimeng Seedance 2 Limited-Time Promotion
  4. 04Mythos Reportedly Breaches Nearly All Pentagon Classified Systems Within Hours
  5. 05Subquadratic Releases SubQ 1.1 Small: Significant Improvement in Attention Computation Efficiency
  6. 06OpenClaw v2026.6.9 Released, Focused on Detail Polish
  7. 07GPT-Image-2 Brand Visual System: From Single Image Generation to Complete Design Suite
  8. 08Nous Research Releases Hermes Agent Blank Slate Mode
  9. 09Xiangyang Qiaomu Open-Sources Codex Icon Generation Skill
06.20Sat · No.81
06.19Fri · No.80
06.18Thu · No.79
06.17Wed · No.78

SpaceX Acquires Cursor and Launches Code Hosting Service, GLM-5.2 Released Open Source, Anthropic’s Frontier Models Designated as National Security Assets

  1. 01SpaceX Acquires Cursor for $60 Billion, Officially Launches Origin Code Hosting Service and 1.5 Trillion Parameter Model
  2. 02GLM-5.2 Officially Released and Open-Sourced, Multiple Bloggers Provide Varied Evaluations
  3. 03WeChat Pay Officially Launches AI-Specific Card, WorkBuddy Among First to Integrate
  4. 04Anthropic's Frontier Models Designated as National Security Assets, UK Citizen Access Requests Denied
  5. 05OpenAI Codex Opens to Third-Party Open-Source Models, Details Three Computer Control Methods
  6. 06Grok Imagine Video 1.5 Fully Launched, Elon Musk Claims Full Film Generation Possible by Year-End
  7. 07NVIDIA ENPIRE Released: AI Agent Autonomously Controls Real Robot for Experiments in Physical World for First Time
  8. 08Midjourney Announces Upcoming Hardware Product Launch
  9. 09Claude Code Founder Shares Agent Engineering Philosophy: Do Less, Let the Model Grow
  10. 10OpenAI GPT-Bidi-1 Voice Model Discovered Early, Microsoft Considers Switching Copilot to DeepSeek
06.16Tue · No.77

SpaceX Acquires Cursor, DeepSeek Completes Special Structure Financing, Anthropic Faces Multiple Crises

  1. 01SpaceX Acquires Cursor in $60 Billion All-Stock Deal, Integrating into SpaceX/xAI Ecosystem
  2. 02DeepSeek Completes 50 Billion Yuan First Round Financing with Rare Structure: Five-Year Lock-Up, Liang Wenfeng's Effective Control
  3. 03Anthropic's Multiple Crises: Pentagon Moves 2/3 of Classified AI Work, Users File Class Action, Opus 4/Sonnet 4 Taken Down
  4. 04Microsoft Hit with Shareholder Class Action, Accused of Concealing Azure Growth Slowdown After $357 Billion Market Cap Loss
  5. 05Mistral New Model Rumors Spark Social Media Frenzy, 'Le Chaton Fat' Tops OpenRouter
  6. 06ByteDance Launches Seedance 2.0 Mini, Reducing Video Generation Costs by About 30%
  7. 07Novo Nordisk Confirms Hack, AI Model Checkpoints and Training Data Stolen
  8. 08Multiple Developer Tool Updates: baoyu-design Supports PPT Export, Papr Lightweight RSS Client Released, Open Design Integrates 150+ Design Systems
06.15Mon · No.76

Anthropic’s Multi-Front Turmoil, Loop Engineering Becomes New Consensus, AMD Launches Local AI Development Platform

  1. 01Anthropic's Multi-Front Turmoil: Fable 5 Removed, KYC Coming, Deception Experiments Spark Deep Discussion
  2. 02"Loop Engineering" Becomes New Consensus in AI Development, Defining Goal Ability Pushed to Core
  3. 03AMD Launches $3,999 Local AI Development Platform with 128GB Unified Memory, Rivaling DGX Spark
  4. 04MiniMax M3 Announces Permanent Half-Price, Cost-Effectiveness Overwhelms Peers
  5. 05Zhipu GLM-5.2 Open-Sourced + Stock Surges 48%; Former Qianwen Lead Lin Junyang's New Lab Valued at $2 Billion
  6. 06SpaceX Market Cap Breaks $2.11 Trillion on Second Day of Listing, Passive Buying Flows In
  7. 07Microsoft CEO Nadella Proposes "Token Capital" Concept, Warns Against Outsourcing Learning Ability in AI Era
06.14Sun · No.75

Alibaba Denies Rumors of Chief Scientist’s Departure, Brazil’s Open-Source Model Accused of Shelling, Anthropic Export Control Details Revealed

  1. 01Alibaba Formally Denies Rumors of Chief Scientist Jingren Zhou's Departure
  2. 02Rio de Janeiro's Open-Source Model Rio 3.5 Accused of Being a Shell of Nex-N2
  3. 03Anthropic Fable 5 Export Control 24-Hour Inside Story: White House and Anthropic Present Conflicting Accounts
  4. 04OpenRouter Launches Fusion: Multi-Model Parallel Fusion to Approach Frontier Models at Low Cost
  5. 05Claude Design's Core Competitiveness Lies in the Model Layer: Harness Engineering Barrier is Low
  6. 06Engineering Trade-offs in Codex's Two Browser Modes: Session Sharing vs. Resource Consumption
  7. 07GLM Widely Used Internally by Major Chinese Tech Firms; Coding Plan API Key Has Whitelist Billing Trap
  8. 08Concentrated Open-Sourcing of AI Agent / Skill Workflows: Covering Writing, Visualization, Programming Containers
06.13Sat · No.74
06.12Fri · No.73
06.11Thu · No.72

Anthropic Fable 5 Sparks Token Cost and Safety Debates, Alibaba Restructures AI Division, Google Open-Sources DiffusionGemma

  1. 01Anthropic Fable 5 Sparks Token Consumption Surge and Safety Policy Controversy
  2. 02Alibaba Shake-Up: DingTalk Leadership Change, Layoff Rumors, and Third AI Restructuring in Three Months
  3. 03ChatGPT Turkey Region Half-Price Offer Ends, AI Subscription Pricing at a Turning Point
  4. 04Google Open-Sources DiffusionGemma: Text Diffusion Paradigm Challenges Autoregressive Monopoly
  5. 05Dario Amodei Publishes Extensive Policy Essay, Defining AI as a National Capability Variable
  6. 06Agent Harness Becomes New Competitive Focus: DeepSeek Hiring, Xiaomi Open-Source, Multiple Frameworks Emerge
  7. 07OpenDoor Cuts 200-Person India Offshore Team, Replaces with AI-Native Small Teams
  8. 08A Developer's Practice Mini-Game Achieves 1 Million UV in 30 Days: An Overseas SEO Case Study
06.10Wed · No.71

Claude Fable 5 release sparks testing frenzy and cost debate, Google launches real-time voice translation

  1. 01Anthropic Releases Claude Fable 5, Multi-Scenario Tests Showcase Astonishing Capabilities
  2. 02Fable 5 Credit Consumption Sparks Debate, Majority of Bloggers Question Its Cost-Effectiveness
  3. 03Fable 5 Hands-On Evaluation: Strong at Vulnerability Analysis but Buggy at Code Writing, UI Design Inferior to 4.8
  4. 04Anthropic's Security Policies Spark Controversy: Distillation Limits, Data Retention, and Partner Cancellations
  5. 05OpenAI CodeX Launches Invite-a-Friend Reset Quota Feature, Joked as Having a Chinese Marketing Style
  6. 06Google Releases Gemini 3.5 Live Translate Real-Time Speech Translation Model
  7. 07Tesla AI6 Chip and FSD Safety Data Improvement in the Netherlands
  8. 08Developer Tools Update: CodePilot, Tokei Fixes, Open-Source Teleprompter
06.09Tue · No.70
06.08Mon · No.69
06.07Sun · No.68
06.06Sat · No.67
06.05Fri · No.66
06.04Thu · No.65

SpaceX sets record IPO valuation at $1.77 trillion, internet bot traffic surpasses humans for first time, Google open-sources Gemma 4 12B

  1. 01SpaceX Files IPO Amendment, Valuation at $1.77 Trillion Sets Record
  2. 02Evan You's VoidZero Team Joins Cloudflare, Vite Ecosystem Gains Platform-Level Support
  3. 03Grok Ecosystem Expands Across Multiple Fronts: Video, Voice, Cloud All Top Rankings
  4. 04Google Open-Sources Gemma 4 12B: Encoder-Free Unified Multimodal Architecture, Runs Locally on 16GB
  5. 05Ideogram 4.0 Open-Sourced: Introduces Bounding Box Precision Control, Text Rendering Reaches Open-Source Ceiling
  6. 06Internet Bot Traffic Surpasses Human Traffic for the First Time: 57.5% of HTML Requests Come from Bots
  7. 07Kimi-Code Rapidly Iterates After Refactoring: Lead Developer Shares a Month of Intense Experience
  8. 08html-video Open-Sourced: Agent Writes HTML to Generate Videos, 30,000 Lines of Code in 3 Days
  9. 09VS Code High-Risk Vulnerability Exposed: Can Steal GitHub Tokens; Security Researcher Discloses Publicly Due to Dissatisfaction with Microsoft
06.03Wed · No.64
06.02Tue · No.63

AI Companies Rush to IPO, WeChat Agent Launch Set for Q3, Doubao Pricing Sparks Debate

  1. 01Anthropic Files S-1 for IPO; MiniMax and Zhipu AI Apply for A-Share Listings
  2. 02WeChat Plans Q3 Launch for Built-in Agent, Deeply Integrating Hunyuan Model
  3. 03ByteDance Seed Core Researcher Quanquan Gu Departs; Previously Rumored to be Pre-training Lead
  4. 04Former Huawei Pangu Model Lead Wang Yunhe Starts AI Agent Company Valued at $100M
  5. 05Doubao Expected to Officially Charge in Late June; Three-Tier Pricing Sparks AI Commercialization Debate
  6. 06Codex and Claude Code Quota Issues Erupt; Claude Fully Resets Limits
  7. 07Cursor Increases Teams User Quotas and Launches Premium Seats
  8. 08Baoyu Releases baoyu-image-gen Skill Supporting Codex CLI Image Generation, Launches New Book "Illustrated Skill"
  9. 09GPT-Image-2 Practical Tip: Use Subtraction to Eliminate the "AI Look"
  10. 10Dreame Technology Initiates Pre-IPO Financing with 70B RMB Valuation
  11. 11OpenAI Models and Codex Now Available on AWS Amazon Bedrock
  12. 12X Launches React with Video Feature, Rivaling Short-Form Video Interaction
06.01Mon · No.62

NVIDIA Launches Three Major Products, MiniMax Unveils M3 Model, VAST Secures Nearly $200 Million in Funding

  1. 01NVIDIA Launches Nemotron 3 Ultra, RTX Spark Superchip, and Cosmos 3 at Computex
  2. 02MiniMax Releases M3 Model: Million-Token Context, Sparse Attention, Native Multimodality
  3. 03VAST Completes Nearly $200 Million Funding, Valuation Reaches $1 Billion
  4. 04Claude Opus 4.8 Receives Polarized Reviews; Multiple Bloggers Share Hands-On Experiences
  5. 05Coze 3.0 Released, Supports Local Agent Integration and Multi-Agent Collaboration
  6. 06Vibe Coding Tool Ecosystem Continues to Explode: Skill Open-Sourced, Codepilot Refactored, Style Atlas Launched
  7. 07OpenAI Voice Hack Night Showcases Agentic Mobile OS Prototype
  8. 08AI Animation and 3D Cultural Tourism Works Enter Mainstream Platforms
  9. 09Unitree Robotics IPO Approved on STAR Market; OpenAI Recruits Robotics Engineers
  10. 10Benedict Evans and Lenny Rachitsky Deep Dive: AI Is at the 1997 Stage

May 20262026 · 31 issuesMay 2026 · Visual Briefs →

05.31Sun · No.61
05.30Sat · No.60
05.29Fri · No.59
05.28Thu · No.58

Cognition Raises $26B Valuation Funding, OpenAI Foundation Invests $250M to Address Economic Impact, X Platform Launches Full Auto-Translation

  1. 01Cognition Raises $1 Billion at $26 Billion Valuation
  2. 02OpenAI Foundation Invests $250 Million to Address AI Economic Impact
  3. 03X Platform Launches Full Auto-Translation, Breaking Multilingual Content Barriers
  4. 04OpenAI to Deprecate Old Models on June 2, GPT-5.5 Becomes Default for Free Users
  5. 05Claude Code Security Guidance Plugin Reaches 160K Installs, Covers Six Risk Types
  6. 06Anthropic Releases Computer Use Best Practices with Counterintuitive Findings
  7. 07MCP Protocol Major Update in July, OpenAI Launches Secure MCP Tunnel
  8. 08AI Industry Enters Reflection Phase: Model+Harness Is the Product
  9. 09Guizang Open-Sources Xiaohongshu Image-Text Layout Skill, Solving AI Content Image Pain Points
  10. 10Xiaomi MiMo Announces 13 Ecosystem Partners, Price Cuts Retain Users After Incentive Program Ends
  11. 11SpaceX Develops Custom C Language AI Training Stack, Outperforming JAX by Over an Order of Magnitude
05.27Wed · No.57
05.26Tue · No.56
05.25Mon · No.55
05.24Sun · No.54
05.23Sat · No.53
05.22Fri · No.52

Codex Lock Screen Remote Control Launches, DeepSeek V4-Pro Permanently Drops Price

  1. 01Codex Thursday Update: Lock Screen Remote Control, Appshots, /goal Officially Launched
  2. 02DeepSeek V4-Pro Announces Permanent Price Cut, Harness Team Begins Mass Hiring
  3. 03Qwen 3.7-Max Released, Outperforms GPT-5.5 and Opus 4.7 in Multiple Benchmarks
  4. 04PwC Paper: Grep Beats Vector Retrieval in Agent Search Scenarios
  5. 05Cloudflare CEO Details: Laying Off 20% of Staff, AI Replaces "Measurers"
  6. 06Microsoft Recalls Internal Claude Code Licenses, Asks Engineers to Switch to Copilot CLI
  7. 07SpaceX Starship V3 First Flight Aborted Due to Hydraulic Pin Failure, Retry Planned for Next Day
  8. 08Xiaohongshu Opens Skill Uploads; Open Design Supports Multi-Project Management and Handoff
  9. 09Open Source Updates: Mega-ASR for Noise Recognition, Google AX Distributed Agent Infrastructure, Feishu Bridge
05.21Thu · No.51

OpenAI Solves Math Conjecture, Google Product Strategy Sparks Trust Crisis, Manus Founders Consider Buyout from Meta

  1. 01OpenAI General Reasoning Model Autonomously Solves Nearly 80-Year-Old Math Conjecture
  2. 02Google AI Product Strategy Triggers Developer Trust Crisis
  3. 03Manus Founders Consider Raising $1 Billion to Buy Back Company from Meta
  4. 04SpaceX Expands AI Computing Services, Partners with Anthropic, and Plans Orbital Data Centers
  5. 05Tesla Discontinues Model S/X and Launches FSD in China
  6. 06Tencent Open-Sources Multilingual Translation Model Hy-MT2, 1.8B Lightweight Version Deployable Locally
  7. 07Hugging Face Open-Sources Ultra-Fast DNA Model Carbon, Processes Complete Human Genome in Two Days on a Single GPU
  8. 08China's AI Short Drama Market Enters Industrial-Scale Production, AI Comic Dramas Become Growth Driver
  9. 09miHoYo Pledges Up to 100 Billion RMB for AI Large Models Over Three Years
  10. 10Independent Developer's Sudden Death Sparks Industry Health Reflection, Vibe Coding Addiction Risk Draws Attention
  11. 11AI Agent Ecosystem Continues to Expand: Grok Build, Hermes Agent, Multica, and More Advance on Multiple Fronts
05.20Wed · No.50

Google I/O Updates, Meta AI Training Controversy, GitHub Extension Breach

  1. 01Google I/O 2026 Unveils Multiple Products, Gemini 3.5 Flash is Fast but Controversial
  2. 02Meta All-Hands Recording Leaked: Zuckerberg Says Company Uses Employees as AI Training Subjects While Cutting 10% of Staff
  3. 03GitHub Hacked: Malicious VS Code Extension Leads to Internal Repository Leak, About 3,800 Repositories Affected
  4. 04DeepSeek Forms Code Team, Recruiting Product Managers and R&D Engineers in Beijing
  5. 05Alibaba Unveils Zhenwu M890 AI Chip and 128-GPU Super Node Server
  6. 06OpenAI Announces Partnership with Singapore Government, Investing S$300 Million to Establish First Overseas Applied AI Lab
  7. 07Cerebras Runs 100-Billion-Parameter Kimi K2.6 at ~1000 tok/s, Setting New Record for Frontier Model Speed
  8. 08Nous Research Releases CNA Method: Steering LLM Behavior Without Training Sparse Autoencoders
  9. 09Railway Banned by Google Cloud, Causing 6-Hour Service Outage
  10. 10OpenAI Offers $2 Million in Tokens to Current YC Batch Startups in Exchange for Equity
  11. 11HuggingFace Launches Hardware Tracking Feature, Showcasing Real Hardware Usage in the Open-Source AI Ecosystem
  12. 12Midjourney Founder Says Using Google TPU for Training May Have Set Research Back a Year
  13. 13Gemini 3.5 Flash Shows Mixed Results in Voice and Task Agent Benchmarks
05.19Tue · No.49
05.18Mon · No.48

Grok Build Integrates Native Image/Video Generation, Anthropic Annualized Revenue Surpasses $30 Billion, AI Disruption Pathways to Middle-Class Formation

  1. 01Grok Build Receives CLI Native Image/Video Generation Capabilities, /imagine-video Enters Open Testing
  2. 02SpaceX Starship Flight 12 Targets Next-Generation Vehicle, Dragon Spacecraft Successfully Docks with ISS
  3. 03Tesla FSD v14.3.3 Released: Smart Summon Speeds Up 33%, False Alerts Significantly Reduced
  4. 04X Platform Officially Surpasses 1 Billion Downloads, User Rating 4.3 Stars
  5. 05The Essence of AI Displacement: Not Jobs, but "Middle-Class Formation Pathways"
  6. 06Eric Schmidt Met with Student Boos at Arizona State University Commencement
  7. 07Anthropic CFO Discloses: Annualized Revenue Surpasses $30 Billion, 90% of Code Written by Claude Code
  8. 08Codex Runner Enables Cross-Machine Control: Distributed Scheduling Practice for an Agent Legion
  9. 09Open-Source Agent IDE ORCA Launches: Supports iOS/Android Clients and Multi-Model Switching
  10. 10Practical Toolset: DeepL API Remains Reliable, WeRead Report Skill Open-Sourced
05.17Sun · No.47
05.16Sat · No.46

OpenAI Consolidates Product Lines to Face Competition, Telecom Enters AI Gateway, Open-Source Attention Mechanism Accelerates Long-Context Processing

  1. 01OpenAI Officially Advances ChatGPT and Codex Merger, Brockman Takes Over Product Strategy
  2. 02Shanghai Telecom Packages Tokens as Phone Plan, 1 Yuan for 250,000 Credit Points
  3. 03NousResearch Open-Sources Lighthouse Attention: Approximately 17x Faster Than Standard Attention at 512K Context
  4. 04Neuralink Showcases First User: Painting Again After Paralysis
  5. 05OpenClaw Engineering Practice: 100 Codex Instances Covering PRs, Code Review, Security, and Automation
  6. 06AI Agent Legion Operations: 20x Cost for 100x Productivity, Agent Network Open-Sourced
  7. 07Codex Cross-Device Control Feature Test: Context Sharing Across Multiple Macs
  8. 08DeepSeek Recruits Agent Harness Product Manager, Defining Model + Harness = Agent
05.15Fri · No.45
05.14Thu · No.44

Claude Code Quota Adjustment, MiniMax Deletion Incident Sparks Security Discussion

  1. 01Claude Code Quota Policy Adjustment: Weekly Limits Relaxed but SDK Usage Capped
  2. 02MiniMax Model File Deletion Incident: User Publicly Reports Abnormal Model Execution and Demands Official Response
  3. 03Nous Research Releases Token Superposition Training: 2-3x Training Efficiency Improvement
  4. 04OpenAI Codex Windows Sandbox Released: 2-Month Free Trial for Enterprises
  5. 05Kimi Launches Web Bridge Browser Extension: Agents Can Automatically Operate Web Pages to Complete Form Tasks
  6. 06Raycast Releases V2 Beta: Upgrades from Launcher to AI Tool Supporting Agents and Skills
  7. 07Discussion on AI-Native Company Characteristics: Long-Term Memory, Workflow Integration, and Experience Accumulation
  8. 08Claude for Small Business Released: Anthropic Integrates Common SaaS Tools like QuickBooks, PayPal
  9. 09OpenSquilla Open-Source Project: Intelligent Model Routing and Local Vector Search Reduce Agent Token Usage
05.13Wed · No.43

Grok Voice Tops Voice Charts, Google and Anthropic Advance Agent Ecosystem

  1. 01Grok Voice Tops Real-Time Voice Charts, xAI Strengthens Voice Interaction Entry Point
  2. 02Google Gemini Enters Android's Core Layer, Cross-App Operation Becomes a System Capability
  3. 03Anthropic Releases Claude for Legal: Legal Plugins, MCP, and Agent Architecture Implemented as a Suite
  4. 04Amazon Racks Up Tokens to Meet KPIs, Big Tech AI Investment Anxiety Surfaces
  5. 05AI Workflow Productization: Value Shifts from Capability to Sellability for Specific Businesses
  6. 06Starship's Twelfth Flight Test Approaches, SpaceX Accelerates Global Spaceport Site Selection
  7. 07WeChat Group Chat One-Click Summary Launched, AI Office Entry Point Continues to Sink
  8. 08Andrew Ng Rebuts AI Job Apocalypse Narrative: Panic Stories Outweigh Real Data
  9. 09Luo Yonghao Joins X, Sparking Discussion in Chinese Twitter Sphere; Lao Luo's IP Becomes a Public Topic Again
05.12Tue · No.42
05.10Sun · No.40
05.09Sat · No.39
05.07Thu · No.37
05.06Wed · No.36

ChatGPT Upgrades to GPT-5.5 Model, Runway Launches Real-Time Conversational Video Characters

  1. 01GPT-5.5 Instant Upgraded to ChatGPT's Default Model, with Multiple Bloggers Discussing Their Experiences
  2. 02OpenAI Launches Migrate to Codex Feature, Supporting One-Click Migration of Other Programming Agent Configurations
  3. 03Runway Launches Characters Feature: Generate a Real-Time Conversational Video Character from a Single Image
  4. 04Google Releases MTP Drafter for Gemma 4, Inference Speed Up to 3x Faster
  5. 05Mininglamp-AI Open-Sources Cider and Mano-P: Turning Mac into a Local AI Workstation
  6. 06Cryptocurrency Exchange Coinbase Announces Layoffs of About 14%, AI Efficiency Cited as One Reason
  7. 07Petdex: An Open-Source Project Providing Third-Party Terminal Pets for Codex Users
  8. 08CodexBar 0.24 and OpenClaw 2026.5.5 Updated Simultaneously
05.05Tue · No.35

Anthropic Releases Jupiter Model, AI Coding Acceleration Layering Sparks Discussion

  1. 01Anthropic Suspected to Release Jupiter Today, OpenAI Codex Sparks Pricing Controversy
  2. 02AI Agent Collaboration Tools Continue to Emerge: OpenClaw Releases Update, Crabbox / gog and Other CLI Tools See Intensive Releases
  3. 03Agent Collaboration Tool Landscape: Moxt / Multica / Slock / Bloome Benchmarked Against Notion / Linear / Slack / WeChat Groups
  4. 04Rethinking the Boundaries of AI Reasoning: Self-Play Can 'Collapse', Adversarial Optimization Must Pair with Independent Discriminators
  5. 05AI Coding Acceleration Layering and Capability Delegation: Frontend >> Backend >> Infrastructure >> Research
  6. 06Knowledge Payment Delivery Dilemma: Students Want to 'Become Idols' Not Knowledge, API Relay Station Customer Acquisition Logic Systematically Dissected
  7. 07Divergence in AI Public Perception: Americans Hold AI Assets Yet Are More Anxious, Chinese Are Optimistic; TV Talk Show "Politically Correct" Attacks AI
  8. 08Content Creation Platform Strategy: Kasha COS Series Premieres on Multiple Platforms, Topaz Annual Subscription Trap Sparks Piracy Discussion
  9. 09Codex /goal Long-Duration Runs: Must Build Acceptance Documents in Phases, Unclear Goals Lead to Agent Early Termination
05.04Mon · No.34
05.03Sun · No.33
05.02Sat · No.32
05.01Fri · No.31

April 20262026 · 30 issuesApril 2026 · Visual Briefs →

04.30Thu · No.30
04.29Wed · No.29
04.28Tue · No.28
04.27Mon · No.27
04.26Sun · No.26

GPT Image 2 Gains Popularity, GPT-5.5 Programming Assistance Sparks Discussion, HTML Slides Emerges as a Trend

  1. 01AI Image Creation Tool Craze: GPT Image 2 Gains Popularity, Seedance Video Generation Clearly Lags Behind
  2. 02GPT-5.5 and New Developments in Programming Tools: Codex Becomes the Hottest Topic
  3. 03Sam Altman: Rethinking Operating Systems and Internet Protocols
  4. 04OpenClaw Ecosystem Remains Active: Record PR Cleanup, Dense Developer Tool Updates
  5. 05Skill.md and Claude.md Documents: The Counter-Consensus Practice That Longer Is Better
  6. 06Rise of HTML Slides: AI Native Companies Have Already Switched, PPT Has Become an Outdated Format
  7. 07Agent Product Design Methodology: PRD No Longer Essential, "Context Seed" Is the Key Design Concept
  8. 08AI Video Models and Vertical Domain Dynamics: Kling 3.0 4K Released, NASA Cargo Spacecraft Launched
  9. 09Elon Musk's Political Speeches: Concentrated Retweets and Discussions
04.25Sat · No.25
04.23Thu · No.23
04.22Wed · No.22
04.21Tue · No.21
04.17Fri · No.17
04.16Thu · No.16

Claude Opus 4.7 Released, GPT Image 2 Realism Raises Concerns, Dreamina Video API Commercialized

  1. 01Claude Opus 4.7 Released, Multiple Bloggers Interpret New Capabilities
  2. 02GPT Image 2 and Mysterious Model "duct-tape" Attract Attention
  3. 03Gemini Mac Client Receives Negative Reviews, Rough Interface Becomes Consensus
  4. 04Dreamina Seedance 2.0 API Launched, AI Video Commercialization Accelerates
  5. 05OpenAI Agents SDK Upgrade: Built-in Sandbox and State Management
  6. 06Skill Essence is Taxonomy, High-Frequency Use Requires Controlling Quantity
  7. 07AI Programming Experience and Agent Capability Boundary Discussion
  8. 08Qwen3.6-35B-A3B Open-Sourced, Elephant Alpha Model Debuts
  9. 09Elon Musk Continues to Focus on South Africa Policy and Starlink
  10. 10Claude Intelligence Downgrade and Token Consumption Issues
  11. 11Other Updates
  12. 12Scraping Statistics
04.14Tue · No.14

Hermes Agent Launches Graphical Interface, Darwin.skill Open-Source Evolution System, Hassabis Predicts AGI Within Five Years

  1. 01Hermes Agent v0.9.0 Release: Dashboard Launches and Explodes on GitHub Trending
  2. 02Uncle Hua Open-Sources Darwin.skill: An Automated Skill Evolution System
  3. 03Gemini + Nano Banana Becomes a Logo Design Powerhouse, Sparking Discussion on Logo Definition
  4. 04Why an AI First Strategy Might Be a Big Mistake?
  5. 05Vercel Open-Sources Open Agents: A Reference Implementation for Enterprise-Level Programming Agent Platform
  6. 06DeepMind Founder Hassabis Biography: Sci-Fi, Power, and Rejecting Facebook
  7. 07Zuckerberg Dialogue Postponed, Hassabis Claims AGI Arriving in Five Years
  8. 08NASA Artemis Follow-up Plans: Moon Landing in 2027, Artemis IV in 2028
  9. 09Anthropic's Guide to Five Multi-Agent Collaboration Patterns
  10. 10The Essence of Large Model Training: Learning is Forgetting, Understanding is Compression
  11. 11Only 17 Tile Patterns: Mathematical Proof of Wallpaper Groups
  12. 12Local Large Model Discussion: Qwen3.5-35B for Code, Gemma 4 for Creative Writing
  13. 13Disabling Claude Code Telemetry Reduces Cache from 1 Hour to 5 Minutes? Anthropic Engineer Clarifies
  14. 14AI Agent Content Creation Efficiency Boost Practice: Cyber Employees + CREAO
  15. 15Elon Musk: Grok Tops Multiple Benchmarks, Tesla US Sales Strong, South African Issues Draw Attention
  16. 16OpenClaw 2026.4.14 Release: GPT-5.4 Routing Optimization & Stability Updates
  17. 17Scraping Statistics
04.13Mon · No.13
04.12Sun · No.12
04.11Sat · No.11
04.10Fri · No.10

Anthropic Advisor Tool Sparks Architecture Discussion, AI Agent Infrastructure and Monetization Tactics in Focus

  1. 01Anthropic Advisor Tool Becomes the Most Focused Product Topic of the Day
  2. 02Codex Subscription Tier Adjustment Also Seen as a Direct Competitive Move Against Anthropic
  3. 03AI Agent Infrastructure and Workflow Discussions Continue to Heat Up
  4. 04Design Skills, Agent Editors, and Installation Methods Become Practical Traffic Points in the Chinese-speaking Sphere
  5. 05Chinese-speaking Sphere Continues High-Frequency Discussion on 'How to Monetize AI Content,' Emphasizing Taste, Conversion, and End-to-End Capabilities
  6. 06AI Video and Digital Human Direction Also Sees a Wave of Product-Side Activity
  7. 07NASA Artemis II Return is the Most Prominent Main Thread Among Non-AI Topics
  8. 08Elon Musk's Account is Very Active, But Topics are Scattered, More of Fragmented Progress on Tesla / X / xAI
  9. 09Crawl Statistics
04.09Thu · No.9

Agent Infrastructure Platformization and Meta’s Multimodal New Route Draw Attention, Skills Tools Workflow Heat Remains High

  1. 01Agent Infrastructure Continues Platformization, Claude Managed Agents Become a High-Frequency Focus
  2. 02After Meta Muse Spark Release, Discussion Focuses on Multimodal Reasoning, Parallel Agents, and Closed-Source Shift
  3. 03Skills and Tools Workflow Remains Very Hot, with Focus on Design, Retrieval, Content Conversion, and Presentation Generation
  4. 04Agent Memory, Context Cost, and Authentication Become Another Main Line
  5. 05Debates on "Distillation/Scaffolding/Product Value" Are Very Intense in Chinese Twitter
  6. 06Elon Musk-Related Content is Abundant, but More Like Single-Account High-Frequency Output, Not Suitable as the Main Conclusion for the Day
  7. 07Aerospace and Exhibition Content Has a Presence, but Belongs to Relatively Independent Branch Lines
04.07Tue · No.7
04.06Mon · No.6
04.05Sun · No.5
04.04Sat · No.4

Claude Third-Party Subscription Ban Sparks Community Shift, Grok Imagine High-Quality Mode Goes Viral, AI Monetization Focuses on Customer Acquisition First

  1. 01Claude Third-Party Subscriptions Blocked, Community Turns to Extra Usage and Alternative Paths
  2. 02Grok Imagine High-Quality Mode Goes Viral, Discussions Focus on Usability and Final Output Quality
  3. 03AI Monetization and Traffic Conversion Continue to Heat Up, Focus on Customer Acquisition First Over Product Development
  4. 04Artemis II Continues as the Main Aerospace Story, NASA Provides Intensive Lunar Mission Updates
  5. 05AI Tools and Workflow Discussions Shift to "How to Find Information and Verify Results"
  6. 06Image Content Creation Circles Review Tutorial Assets and Emphasize "Post-Processing Over Pure Prompting"
  7. 07AI Discussion Heat in Offline Scenarios is Also Penetrating Mass Professional Groups
04.03Fri · No.3
04.02Thu · No.2
04.01Wed · No.1