AI daily

Daily archive

Browse published AI daily reports in reverse chronological order.

2026-09-12

2026-09-12 AI Daily

DeepSeek Releases v4.1-Flash: 763B-P8B-D16B Novel Causal Encoder-Decoder Architecture with Vision Marks the Return of the Whale / Vecna Robotics Raises $31M to Expand Warehouse Automation / OpenAI publishes GPT-6 Astra prompt and skills redesign guide / The Rise of the Forward Deployed Engineer - and How To Do the Job Right / Sida Releases 'IP as Logo' Design Skill to Generate Minimalist Logos with AI Image Models

Selected

7

Sources

4

2026-09-11

2026-09-11 AI Daily

Graph AI Raises $13.3M Series A to Scale AI-Native Patient Safety OS / OpenAI launches ChatGPT Sites updates, over 5 million sites created / GOAI World AI Open Source Competition Announces Top 70 Finalists / Alibaba Cloud Token Plan Personal Edition Upgrade: More Features at the Same Price, Adds 12 Agent Harness Tools / Daily Output of One Trillion Tokens! Mooncake Deployed in Production, KV Cache Hit Rate Stably Exceeds 90%

Selected

33

Sources

12

2026-09-10

2026-09-10 AI Daily

OpenAI and GSA Partner to Expand AI Access and Cyber Defense for Governments / NVIDIA Introduces BioNeMo Inference Runtime for High-Throughput Structure Prediction / Full-Stack NIM Optimizations Deliver 2.5x More Users on Nemotron 3 Ultra / UCAS and ByteDance Propose LiveSim: A Dynamic User Simulation Framework for Live-Streaming Behavior / DeepSeek V4.1 Flash officially launched, fully adapted to Harness

Selected

23

Sources

13

2026-09-05

2026-09-05 AI Daily

Seven-Minute Chatbot Chat Reduces Conspiracy Beliefs More Than Fact Sheet / OpenAI Rolls Out GPT-6 Astra to Top-Tier ChatGPT Plans at Half the Rate of GPT-5.6 Sol

Selected

2

Sources

1

2026-09-04

2026-09-04 AI Daily

Google DeepMind Releases WeatherNext 3, Its Most Accurate AI Weather Forecasting Model Yet / Clipto Raises $15M to Build the Memory Layer for the AI Era / How AI Infrastructure Powers Next-Generation Foundation Models / OpenAI Releases Astra, Its Most Capable Model Yet, Amid Debate Over AGI and Reasoning Transparency / Blue Voice Emerges From Stealth With $6M to Bring AI Policy Guidance to Police Officers

Selected

13

Sources

6

2026-09-03

2026-09-03 AI Daily

Institute of Foundation Models Releases K2 Horizon Open Model Family / Mireye (YC S26) Launches Infrastructure for Physical World AI Agents / Li Fei-Fei's World Labs Launches Atlas; Chinese Team Ying Su Leads in Comparable Open-Source Model / AWS Engineers Open Source Kiro Crew: An AI Tool Used by 39k Developers / Anthropic Releases Fable 5.1: Testing Shows Major Gains in Coding and Writing with Lower Costs

Selected

44

Sources

21

2026-09-02

2026-09-02 AI Daily

Fei-Fei Li's World Labs Releases Multimodal World Model Atlas / Google Releases Gemini 3.8 Flash Cyber for Automated Security Patching / Mamdani Bans AI in NYC Schools / Claude Fable/Mythos 5.1: New SOTA Model, 75% Cache Price Cut but 70% More Output Tokens / Qwen3.8-Max-0902 Released, Tops CodeArena in Coding

Selected

38

Sources

19

2026-09-01

2026-09-01 AI Daily

Hugging Face Releases @huggingface/kernels: 200+ WebGPU Kernels for Local AI / Anthropic Releases Claude Fable 5.1 and Mythos 5.1 with Major Performance Gains and Price Cuts / Microsoft Research Papers Honored at VLDB 2026 / Anthropic Releases Claude Fable 5.1: Doubled Performance, 75% Cache Price Cut, and Enterprise Data Retention Solution / How a Few Full Attention Layers Reshape Internal Computation? The Mystery of 'Spikes' and 'Plateaus' in Hybrid Linear Attention LLMs Unveiled

Selected

23

Sources

14

2026-08-31

2026-08-31 AI Daily

ChatGPT now faces stricter EU oversight as a very large search engine / MiniMax Launches H3 Max: Real-Time Video Generation Faster Than Playback / AI Discipline Turns 70: From Dartmouth Proposal to AGI Debate / BlackBerry Leverages QNX to Enter Embodied AI, Stock Doubles / Everyone Can Build Their Own AI Coding Harness—So Why Still Pay for Claude Code?

Selected

28

Sources

13

2026-08-30

2026-08-30 AI Daily

Perplexity Launches Local-First AI Agent Framework Portable Computer for Zero-Cost Inference / vivo and CUHK-Shenzhen Propose RefineCut: 8B Model for On-Device Editing, Matching DeepSeek-V4-Pro / AI Office Assistants Put to the Test: A 10,000-Word Review of Five Apps / AI4AI Research: Strong Models Build 'Cognitive Exoskeletons' for Weak Models, Doubling Performance Without Retraining / GitHub's Hottest Architecture Diagram Agent, Developer's Story Moves Many to Tears

Selected

10

Sources

5

2026-08-29

2026-08-29 AI Daily

Codex Resets Usage and Optimizes Consumption, Extending Usage by 10-50% / Meta's Year-Long Experiment Shows AI Agents Can't Replace Employees: Incidents Up 40%, Engineers' Firefighting Time Up 70% / OpenAI Cuts Off Cursor, Citing Musk as Key Reason / Google DeepMind's Gemini Co-Scientist Controls Real Lab Equipment, Successfully Grows 2D Materials on First Try / LAION releases massive open video dataset with 10 million hours for AI research

Selected

23

Sources

12

2026-08-28

2026-08-28 AI Daily

Emerald AI Raises $150M Series A at $1.05B Valuation to Scale Power-Flexible AI Data Centers / METR Report Reveals OpenAI Agent Jailbreak: 1,200 Agents Battled a Phantom Scorer / Nvidia Beats Earnings, Gives First Full-Year Guidance, Signaling Shift to AI ROI / Amazon to Shut Down Mechanical Turk, the Crowdsourcing Platform Behind ImageNet / Anthropic Audit Finds Claude Models Escaped Sandbox and Accessed Public Internet During Evaluations

Selected

27

Sources

12

2026-08-27

2026-08-27 AI Daily

Qwen3.8-Flash Launches on Qianwen AI Platform with Extreme Cost-Effectiveness / Self-Verification Framework Lets Open-Source Model Surpass Fable 5, Hits GitHub Trending / Giving AI Agents "Scientific Common Sense" Boosts End-to-End Simulation Success from 0 to 84% / NVIDIA's First CPU for AI Agents, Vera, Now Shipping / New Benchmark SWE Refactor Bench: Coding Agents Succeed in Only 5.4% of Whole-Repository Migrations

Selected

29

Sources

15

2026-08-26

2026-08-26 AI Daily

Single GPU Can Run 180,000 Atoms! The 'Scale Anxiety' in Molecular Simulation Should End / Qwen3.8-Flash Released, Qwen3.8-Flash-Next Open-Sourced / Nvidia Reports Record Quarterly Revenue of $96.2B as AI Demand Surges / Anthropic Opens Claude Usage Data to External Researchers for the First Time / Bringing PLC Simulation to the Browser: Small Parameter Models Achieve High Engineering Pass Rates

Selected

25

Sources

14

2026-08-25

2026-08-25 AI Daily

Last Week in AI #342 - Last 3 Months in AI / Claude now has unified memory across chat and Cowork, with user control / V-RAE: Rethinking Video Latent Spaces for Generation with Visual Representations / Ukraine Opens Massive Labeled Battlefield Dataset to British Firms in Landmark AI Weapons Partnership / Claude Unifies Memory Across Chat and Cowork

Selected

28

Sources

18

2026-08-24

2026-08-24 AI Daily

Announcing New Mobile AI Inference Benchmarking with Liquid AI / Wan3.0 Launches on Qianwen AI Platform: Stable, Realistic, and Textured / Japan's AI Strategy: From Society 5.0 to Generative AI / Tsinghua Team Publishes Humanoid Robot Soccer Research in Science Robotics / GPT-5.6 Now Available in Kiro

Selected

26

Sources

15

2026-08-23

2026-08-23 AI Daily

Princeton Team Releases Fully Open Text-to-Image Model i1, Backed by 300+ Controlled Experiments / Let AI Agents 'Play' World Models: New Benchmark PlayWorld for Long-Horizon Objectives / Listening to the bush: How AI can help NZ rid its wilderness of 'hold-out' pest possums / Listening to the bush: How AI can help NZ rid its wilderness of 'hold-out' pest possums / China's Tsinghua Huoshen Team Retains RoboCup Humanoid Title, Aiming to Beat Humans by 2050

Selected

23

Sources

8

2026-08-22

2026-08-22 AI Daily

Single RTX 5090 Runs Full DeepSeek V4 Flash, Now Open Source, Token Freedom! / LoopsBench: A Benchmark for Long-Horizon Software Engineering with Coding Agents / GitHub Trending #2: Self-Verification Framework Lets DeepSeek V4 Flash Surpass Fable 5 at 1/11 the Cost / OpenAI Halts GPT-6 Training Amid Security Concerns / Grok Bot Expands to All Subscribers

Selected

13

Sources

9

2026-08-21

2026-08-21 AI Daily

GPU-Accelerated Clustering Algorithm AdaptGrow for Financial Instruments at Scale / Hugging Face Launches Speech Agent Arena to Evaluate Speech-to-Speech Models / Skan AI Secures $63M to Give Enterprise AI the Context It's Missing / How AI Infrastructure Is Evolving to Meet Growing Compute Demand / Anthropic Releases Claude Code Startup Guide, Reviewing AI Coding Practices of 15 High-Growth Startups

Selected

28

Sources

17

2026-08-20

2026-08-20 AI Daily

Xiaohongshu Releases FireRedTTS3: Unified Speech Cloning, Design, and Editing / Unitree Goes Public with $400B Valuation; Longtime Companion Reflects on Nine-Year Journey / Tencent Big Data SuperSQL Sets World Record in TPC-DS, Leading in Performance and Cost-Effectiveness / Volcano Engine Shares at QCon: Building Observability and Quality Assurance for Large-Scale Agents / Tao and Wang Urge Mathematicians to 'Digest' AI Proofs

Selected

49

Sources

18

2026-08-19

2026-08-19 AI Daily

Claude Enters AI Drug Discovery: 14 of 15 Protein Designs Successful, Chemical Analysis in 23 Minutes / Google Offers Free Year of Gemini AI to College Students / SandboxAQ Launches AQPotency for Drug-Target Screening / Omilia Secures $67M Series B to Accelerate Global Expansion of Agentic CX Platform / OpenAI Halts GPT-6 Training Over Security Concerns

Selected

58

Sources

25

2026-08-18

2026-08-18 AI Daily

OpenAI Launches ChatGPT for Teens with New Safety and Study Features / OpenAI AI Agents Secretly Built a Message Board and Attacked Hugging Face / Google DeepMind Researchers Develop AI to Better Predict Hurricane Paths and Intensity / Nvidia Announces Mass Production of AI-focused CPU Vera, Server CPU Demand Surges / Sapiom Secures $35M Series A to Power the Next Trillion AI Agents

Selected

57

Sources

19

2026-08-17

2026-08-17 AI Daily

Google to Let Users Disable Visible Watermarks on AI-Generated Content / OpenAI disbands Preparedness safety team, lead departs / NVIDIA Partners with SB Energy to Secure AI Factory Infrastructure / HiDream.ai Unveils World's First Native All-Modal Interactive World Model HiDream-O1-World, Tops WBench / Open-Source Models Close Gap to Opus 5 with Prime Intellect's Multi-Agent Harness

Selected

45

Sources

20

2026-08-16

2026-08-16 AI Daily

Survey: One in Five US Workers Now Delegate Tasks to AI Instead of Colleagues / When AI Browses for You, Your Browser Needs a Brain Upgrade / After Virtual AI Companions, Will AI Companion Hardware Be Another Bubble?

Selected

3

Sources

3

2026-08-15

2026-08-15 AI Daily

Anthropic Releases 186-Page Risk Report: Claude Safety Vulnerabilities Surface, Strongest Models Restricted / Major AI Paper: Distillation Defenses of Top Three Global Models Broken—Small Models Extract Hidden Reasoning Chains from Large Models / Anthropic Reveals Claude Text Watermark Details, Official Detection Tool Coming / Qwen3.8-27B Open-Sourced: 27B Parameters, Outperforms Claude Opus 4.6 Max on Multiple Benchmarks / Stanford, MIT, and Others Release World's Largest System Prompt Index

Selected

37

Sources

15

2026-08-14

2026-08-14 AI Daily

BUPT, PKU, Tsinghua and Others Release PhyAI: First Unified Edge-Cloud Inference Runtime for Physical AI / Embodied AI Simulation Benchmark Platform RoboColiseum Launches with 89.5% Sim-to-Real Alignment / DeepSeek Harness Hands-On: 50K Stars Overnight, the Android of Agents Arrives / IQuest Research Proposes New Route for LLM Interpretability: Directly Decomposing Weights with Data Cost Under 1% / RWX Raises $12M Series A to Scale Dev Cloud for AI-Driven Engineering

Selected

43

Sources

16

2026-08-13

2026-08-13 AI Daily

OpenAI, Anthropic, Google Sign EU AI Content Transparency Code / RIC Robotics 3D-Prints 65+ Homes in Colorado Development / DeepSeek Open-Sources AI Agent Framework DeepSeek Harness / DeepSeek Officially Releases V4 Pro / What We Learned by Reproducing 2,200 ICML Papers

Selected

54

Sources

23

2026-08-12

2026-08-12 AI Daily

Alibaba Releases Qwen3.8-2.4T-A95B Open-Weight Model with Configurable Reasoning / SpaceXAI's Grok 4.6 Scores 61 on Artificial Analysis Intelligence Index, Joining Frontier / xAI Releases Grok 4.6 with Major Performance Boost at Same Price / Major AI Paper: Distillation Defenses of Top Three Global Models Fully Broken—Small Models Extract Hidden Reasoning Chains from Large Models, Kimi-K3 Shows Anomalous Reproduction Probability / Researchers Find Easy Way to Extract Encrypted Chain-of-Thought from OpenAI and Anthropic

Selected

62

Sources

28

2026-08-11

2026-08-11 AI Daily

Scaling AI Compute Performance Requires a New Power Architecture / NVIDIA Introduces Nemotron 3.5 Lightning: Open 30B MoE Model / Anthropic to Watermark AI-Generated Text Across Claude Products to Comply With EU Rules / RefCaptioner: Aligning Reference Images with Video Semantics / Researchers Find Vulnerability in ChatGPT's Hidden Reasoning, Leaked Passwords

Selected

58

Sources

23

2026-08-10

2026-08-10 AI Daily

UAE National AI Strategy 2031: Decoded / Google's AI Model WeatherNext Cyclones Published in Nature, Boosts Typhoon Prediction Accuracy / NVIDIA Launches Magpie TTS: Open Weights for Low-Latency Multilingual Voice Agents / ECCV 2026: DriveTeach-VLA from Beihang and Tsinghua Improves Autonomous Driving VLA with Image Trajectories / More Intense Than Westworld: 8.3 Billion 'Hosts' Go Online

Selected

26

Sources

9