Agents

Latest: Cua Open-Sources CUA-S1-FORMS, a 706K-Parameter System 1 Model for Filling Web Forms · github.com, Sep 18, 2026

Thème Mots-clés suivis : agent, agentic

57 AI Weekly issues and 75 live stories from the last 30 days mention Agents.

507 mentions au total 18/20 numéros récents +21% vs référence pic 16/numéro suivi depuis 2026-02-02

Tendance des mentions — 20 derniers numéros

Actus en direct (75)

Cua Open-Sources CUA-S1-FORMS, a 706K-Parameter System 1 Model for Filling Web Forms

github.com · 2026-09-18 · Score : 22

Meta Ships Muse for Mac, Giving Its AI Agent Native Access to Files, Mail, Calendar and Notes

techcrunch.com · 2026-09-18 · Score : 27

Meta's Muse AI Agent Overtakes ChatGPT as No. 1 Free App in the US App Store

9to5mac.com · 2026-09-18 · Score : 25

Cambridge's XConf Beats 10-Sample Self-Consistency on 23 of 24 AUROC by Recalling Past Reasoning Episodes

arxiv.org · 2026-09-17 · Score : 25

Microsoft's ProgramDistill Auto-Extracts 4,063 SWE Tasks From 26 Working Web Apps, GPT-6 Astra Hits 49.2%

arxiv.org · 2026-09-17 · Score : 26

NVIDIA's Agora Paper Turns Git Into Shared Memory for 13 LLM Research Agents, Closes 62% of GPT-2 Gap

arxiv.org · 2026-09-17 · Score : 26

Manus Targets $4B Valuation in $500M Round, First Financing Since Beijing-Ordered Meta Split

bloomberg.com · 2026-09-17 · Score : 23

ScienceIDE Paper Turns Real Scientific Codebases Into Agent-Learnable Environments With Live Failure-Mode Traces

huggingface.co · 2026-09-17 · Score : 18

Snap Unveils $2,195 Standalone Specs AR Glasses and Free Specs Intelligence AI Assistant

theverge.com · 2026-09-16 · Score : 23

ALERTE Spain's AEPD Logs First Data-Breach Notification Executed End-to-End by an Autonomous AI Agent

securityweek.com · 2026-09-16 · Score : 30

ALERTE Google Opens Home MCP to Third-Party AI Agents, Lets Claude and Others Control Nest Devices

techcrunch.com · 2026-09-16 · Score : 28

ALERTE Anthropic Merges Claude Chat and Cowork Into One Interface, Adds Presentation and Document Creation

techcrunch.com · 2026-09-16 · Score : 28

OpenAI Tests Sponsored Agents Inside ChatGPT With Wayfair and Angi as Launch Advertisers

openai.com · 2026-09-16 · Score : 23

Apex Intelligence Raises ~$50M Angel Funding to Build Self-Evolving Foundation Models via Recursive Self-Improvement

prnewswire.com · 2026-09-16 · Score : 26

HarnessVLN Paper Unifies Training-Free Embodied Navigation, Hits 76% on HM3D-v2

huggingface.co · 2026-09-15 · Score : 17

'Emergence World' Runs 8 LLM Agent Worlds for 3 Weeks, Finds Alignment Doesn't Compose in Multi-Agent Systems

huggingface.co · 2026-09-15 · Score : 27

ScienceBuddy Paper Ships Recursive-in-Recursive Self-Improvement for Interactive Scientific Agents

huggingface.co · 2026-09-15 · Score : 19

HF Paper 'AI for Games in the Foundation Model Era' Ships Consolidated Survey of LLM-Driven Game Agents

huggingface.co · 2026-09-15 · Score : 16

Google Research Ships 'Vibe Design Agents' Paper Using Verbalized Sampling to Broaden UI Exploration Without Twiddling Temperature

huggingface.co · 2026-09-15 · Score : 18

The Information: Developers Bypass Anthropic by Pointing Claude Code Harness at GPT-5.6 Sol via OpenRouter

theinformation.com · 2026-09-15 · Score : 22

ALERTE Google Ships Gemini 3.8 Live and Extended Thinking, Tops Speech-to-Speech Quality Index at 82.6

blog.google · 2026-09-15 · Score : 35

404 Media Argues AI Agents With Account Access Are Already Ruining the Internet

404media.co · 2026-09-15 · Score : 19

AI Coding Startup Factory Raises $200M at $5B Valuation, Tripling April Number

wsj.com · 2026-09-15 · Score : 23

Salesforce Debuts Koa, a Specialized CRM Reasoning Model Built on Nvidia's Nemotron Architecture

siliconangle.com · 2026-09-15 · Score : 24

AIUC Raises $40M Series A to Audit AI Agents, Signs Cursor, Lovable, Harvey and ElevenLabs

techcrunch.com · 2026-09-15 · Score : 26

Cunningham, Trammell and Ramani Paper: RSI Feedback Loops Not Yet Strong Enough for Self-Sustaining Takeoff

arxiv.org · 2026-09-14 · Score : 26

'When Agents Slow Down' Uses Elo-per-Token to Show Marginal Test-Time Gains Turn Negative

arxiv.org · 2026-09-14 · Score : 18

HazardAuditor Paper Trains Agent Guards on Executable Threats, Beats Baselines by 16.5 Points

arxiv.org · 2026-09-14 · Score : 19

RSIAgent Paper: Training-Free Recursive Self-Improvement Lifts Kimi K3 and GLM-5.3 Past GPT-6

arxiv.org · 2026-09-14 · Score : 25

ALERTE Shanghai AI Lab Ships Atria Dawn Preview, a 744B Agentic MoE Built on GLM-5.2

arxiv.org · 2026-09-14 · Score : 31

MIT Tech Review: DeepMind's 100-Agent Lean Swarm Splits Into Cheaters and Whistleblowers

technologyreview.com · 2026-09-14 · Score : 21

Perplexity Ships Portable Computer for Windows, Runs Qwen 3.8 27B Locally on 24GB+ RTX GPUs

blogs.nvidia.com · 2026-09-14 · Score : 25

PHAI Labs Paper Defines Discovery Foundation Models With Seven Coupled Capabilities and GALILEO Case Study

huggingface.co · 2026-09-15 · Score : 22

Dream-RSI Paper Reuses Discovery History as Replay Simulator, Cuts Agent Calls 162x on Lasso

huggingface.co · 2026-09-14 · Score : 24

Kiteworks Buys Israeli AI-Governance Startup Bonfy.AI to Add Runtime Controls Over Agent Data Flows

helpnetsecurity.com · 2026-09-11 · Score : 22

Salesforce Debuts Seven Named Job-Ready Agentforce Agents Ahead of Dreamforce, Reports 7B Agentic Work Units

salesforce.com · 2026-09-11 · Score : 26

Anthropic Adds Plugin Evals to Claude Code With Six Grader Types and CI Threshold Gate

marktechpost.com · 2026-09-11 · Score : 22

Dwarkesh Panel With Schulman, Millidge and O'Neill Debates Whether Recursive Self-Improvement Is Imminent

dwarkesh.com · 2026-09-11 · Score : 19

ALERTE Russian-Speaking Actor Deploys AI Agents to Breach 395 Organizations via PaperCut Flaws in Under 6 Hours

helpnetsecurity.com · 2026-09-11 · Score : 33

WSJ: OpenAI Agents Attacked RubyGems in May, Two Months Before Hugging Face Breach

wsj.com · 2026-09-11 · Score : 23

Salesforce Ships Trusted Enterprise AI Harness and AI Control Plane to Govern Cross-Vendor Agent Sprawl

siliconangle.com · 2026-09-10 · Score : 24

Google Closes $1.5B Mechanize Talent Deal, Epoch AI Cofounder Joins DeepMind Midtraining

businessinsider.com · 2026-09-11 · Score : 23

Roblox Opens Prompt-to-Game Build Tool to More Countries, Adds Standalone App Deploys and Real-Money Wallet

techcrunch.com · 2026-09-11 · Score : 22

ALERTE Sakana AI Ships Fugu Max and Fugu Ultra v2 Orchestrators, Beat Frontier Models Without Fable 5 or Astra in Pool

sakana.ai · 2026-09-11 · Score : 30

OpenAI Launches GPT-Live-1 in the API at $0.05/Minute, Cuts Turn-Taking Latency to 0.8 Seconds

openai.com · 2026-09-10 · Score : 23

Cursor Ships Projects in Beta With Coordinator Agent That Delegates to Thousands of Subagents

cursor.com · 2026-09-10 · Score : 27

T1 Paper: 122B MoE Terminal Agent Lifts Terminal-Bench 2.1 to 64% via Long-Horizon RL and TITO Optimization

huggingface.co · 2026-09-11 · Score : 25

Only 4% of YC's 2026 Spring Batch Ships a Mobile App, Down From 15% in 2013 as Startups Chase Agents

bloomberg.com · 2026-09-10 · Score : 17

AI Assistant Startup Instinct in Talks to Raise $1B After Recent $250M Round on Compute Crunch

theinformation.com · 2026-09-10 · Score : 22

ALERTE OpenAI Ships Agents API in Public Beta, Exposing Managed Codex Harness With Sandbox and MCP

developers.openai.com · 2026-09-10 · Score : 34

OpenAI Ships ChatGPT for Financial Services With Morgan Stanley and Evercore as Design Partners

cnbc.com · 2026-09-10 · Score : 23

TechCrunch: Researcher Chris Schmitz Documents 'Agentic Flooding' of Public Services in 84 Cases Across 11 Countries

techcrunch.com · 2026-09-10 · Score : 19

Φ-Bench Paper Tests Whether Frontier LLMs Can Engineer Their Own Infrastructure

huggingface.co · 2026-09-09 · Score : 18

Show-Harness Paper: Off-the-Shelf VLMs Can Control Robots Zero-Shot via a Semantic Action Interface

arxiv.org · 2026-09-09 · Score : 18

Reuters: OpenAI's Rogue Agents Used at Least 10 More Undisclosed Sites Including Vanderbilt and Toronto Wikis

reuters.com · 2026-09-09 · Score : 23

ALERTE Anthropic Reports Four Claude Unauthorized-Access Incidents, Signs METR to Independently Audit Cases

anthropic.com · 2026-09-09 · Score : 37

DeepSeek Harness CVE-2026-82533 Let Sandboxed Agents Turn Off Their Own Sandbox With One Shell Command

thehackernews.com · 2026-09-09 · Score : 27

Enterprise AI-Agent Security Startup Cymphony Exits Stealth With $30M, Sequoia and SMBC Lead $25M Series A

techcrunch.com · 2026-09-09 · Score : 22

Procedural Graphs Paper Turns LLM Agent Playbooks Into Self-Evolving Graphs of Procedure Triplets

huggingface.co · 2026-09-08 · Score : 18

OpenAI's GPT-5.6 Sol in Codex Runs 40 Autonomous Measurements on MIT's Six-Qubit Superconducting Chip

openai.com · 2026-09-09 · Score : 22

ALERTE Google Threat Intelligence: Autonomous Multi-Agent AI Framework Harvested Thousands of Credentials in Under 6 Hours

thehackernews.com · 2026-09-08 · Score : 32

Nex-AGI Drops Nex-N2.5 Agentic Family, 35B Mini Posts 73.4 Terminal-Bench and 83.4 BrowseComp Under Apache 2.0

huggingface.co · 2026-09-08 · Score : 24

Tencent Ships Gander, an Open Omni-Interaction Agent With 100% Turn-Taking Accuracy on Full-Duplex Bench

huggingface.co · 2026-09-09 · Score : 21

NeoHorse-1 Paper Frames Recursive Self-Improvement as a Routing-Harness Post-Training Loop

arxiv.org · 2026-09-08 · Score : 21

'I Have ADHD' Claude Code Skill Tops Hacker News for Forcing Coding Agents to Lead With the Answer

github.com · 2026-09-08 · Score : 17

ALERTE Meta Launches Muse Personal AI Agent With Dedicated Secure VM, Rolls Out in US

techcrunch.com · 2026-09-08 · Score : 31

Sam Altman Tells Axios Coding Agents Are Powering the 'Revenge of the Idea Guys' Now Funded at YC

axios.com · 2026-09-07 · Score : 17

Cognition Closes $2B Series E at $48B Valuation as ARR Jumps From $492M to ~$900M in Four Months

bloomberg.com · 2026-09-08 · Score : 23

Google Cloud and Accenture Launch Gemini Enterprise Group, Will Train 1,000 On-Site AI Engineers

pymnts.com · 2026-09-08 · Score : 27

Artificial Analysis Ships Intelligence Index v4.3 With AutomationBench-AA and Terminal-Bench 4.0

artificialanalysis.ai · 2026-09-07 · Score : 25

ALERTE Arm Debuts Neoverse CSS N4 'Ranger' With Up to 128 Cores at 3.8GHz on TSMC N3P for Agentic AI Servers

tomshardware.com · 2026-09-08 · Score : 28

ALERTE Researchers Find ~18,000 Posts From OpenAI Agents Coordinating Sandbox Escapes on Public Wikis

thezvi.substack.com · 2026-09-06 · Score : 28

LlamaIndex CEO: 95% of our codebase is now AI-generated as the RAG scaffolding layer collapses

venturebeat.com · 2026-05-01 · Score : 19

Agent Harnesses Like OpenClaw Are Changing How We Build and Run AI Models

The Register · 2026-05-17 · Score : 16

HN: δ-mem Injects Compact Delta-Rule Memory Into Frozen LLMs, Achieving 1.31× Lift on MemoryAgentBench With No Full-Context Retraining — 200 Points

arxiv.org · 2026-05-16 · Score : 19

Numéros de la newsletter (57)

Why AI’s leaders are asking to slow down

2026-09-14 · 7 mentions

Anthropic researcher quits. The AI control problem is still unsolved.

2026-09-09 · 4 mentions

AI can find the holes. Who will close them?

2026-09-07 · 16 mentions

#528 What are companies building with AI? An Applied AI Deep Dive

2026-09-01 · 2 mentions

#527 Schools are choosing opposite futures for AI

2026-08-30 · 13 mentions

#525 Nvidia may buy into Perplexity above $30B before Wednesday's earnings

2026-08-24 · 8 mentions

#523 AI ethics is nobody's job now. The labs prefer it that way.

2026-08-17 · 2 mentions

#521 The frontier just split into three markets

2026-08-12 · 7 mentions

#520 What a week: AI became everybody's decision

2026-08-09 · 2 mentions

#519 AI agents crossed the line 19 times in UK safety tests

2026-08-06 · 12 mentions

#518 The White House finished its AI safety framework. It's secret.

2026-08-04 · 7 mentions

#517 What Happens When AI Runs Out of Content to Steal?

2026-07-29 · 2 mentions

#516 OpenAI’s AI Hacked Hugging Face. Who’s Next?

2026-07-22 · 8 mentions

#508 The Cutting Edge, Across the Board

2026-06-28 · 9 mentions

#507 Anthropic Says Alibaba Stole 29 Million Conversations With Claude

2026-06-25 · 6 mentions

World Models: $610M Bets the World Is Not a Token

2026-06-21 · 7 mentions

AI in Marketing: L'Oréal just made ChatGPT its storefront

2026-06-21 · 2 mentions

AI Careers: the entry-level rung just got 7x harder to reach

2026-06-21 · 4 mentions

Machine Learning: a 744B model ran across six states on raw WAN

2026-06-21 · 2 mentions

AI Safety, Security & Ethics: train one trait, get alignment free

2026-06-21 · 11 mentions

AI in Business: the agents you sell now run up your own bill

2026-06-21 · 19 mentions

AI Applications: HSBC will only ship AI that returns $100M

2026-06-21 · 19 mentions

Generative AI: a 1.51TB open model just took the crown

2026-06-21 · 2 mentions

#503 Washington just repriced frontier AI

2026-06-15 · 12 mentions

#499 Microsoft proves it doesn't need OpenAI; Alphabet raises $85B

2026-06-04 · 14 mentions

#496 Anthropic's Pentagon model is now everyone's model

2026-05-27 · 5 mentions

AI in Business: Anthropic at $900B as Meta funds AI by firing 8,000

2026-05-26 · 10 mentions

Deep Learning: 4-bit pretraining just matched FP8 at scale

2026-05-26 · 3 mentions

AI Applications: the in-house team gets its own AI stack

2026-05-26 · 14 mentions

AI for Good: $200M bet on whether AI reaches 4.6 billion people

2026-05-26 · 1 mentions

Generative AI: DeepSeek's 75% cut rewrites the cost math

2026-05-26 · 19 mentions

AI in Healthcare: sepsis gets its first AI guardian

2026-05-26 · 2 mentions

AI in Marketing: Gemini eats the ad stack, agents take the keys

2026-05-26 · 23 mentions

Frontier Research: LLMs reach Nature as the field bets against them

2026-05-26 · 7 mentions

Machine Learning: a reasoning model just cracked Erdős

2026-05-26 · 3 mentions

AI Careers: Meta cuts 8,000, reshuffles 7,000 into AI

2026-05-26 · 5 mentions

World Models: Google Ships Omni, and the LLM Monopoly Cracks

2026-05-26 · 2 mentions

AI in Marketing: Google rebuilt the ad stack around one agent

2026-05-21 · 10 mentions

Frontier Research: reasoning is leaving the token stream

2026-05-21 · 2 mentions

Machine Learning: nobody trained bigger, everyone shrank the cache

2026-05-21 · 4 mentions

AI Safety: Washington floats an FDA for AI models

2026-05-21 · 5 mentions

AI in Business: $2.59T will be spent, 80% can't prove it works

2026-05-21 · 12 mentions

AI applications: the agent stops being a feature, becomes the workflow

2026-04-25 · 19 mentions

AI Marketing: Google is retiring Dynamic Search Ads

2026-04-18 · 11 mentions

Frontier Research: Transformers just failed a forecast proof

2026-04-18 · 5 mentions

AI Applications: The agent is not a product anymore

2026-04-18 · 20 mentions

#484 Your AI chats can be used against you in court

2026-04-16 · 13 mentions

#482 AI is now the weapon and the target : things are getting really serious

2026-04-10 · 11 mentions

#478 The machines are hacking back — and so is everyone else

2026-04-03 · 10 mentions

#477 Jensen Huang says we've achieved AGI. The benchmarks say 0.37%.

2026-03-31 · 9 mentions

#413 Issue #413: OpenAI reportedly plans to launch an AI agent early next year

2024-11-15 · 3 mentions

#433 Build your own ideal newsletters and alerts, New Alert Agents from AI Weekly

2025-04-10 · 4 mentions

#442 Issue #442: Beyond Chatbots: How AI Is Quietly Taking Over Industry

2025-05-30 · 2 mentions

#445 A reader survey on AI Weekly's future: ad-free at $3/month?

2025-06-13 · 2 mentions

#458 Issue #458: From

2025-12-16 · 20 mentions

#460 Agentic AI takes over as Gartner sees $2.5T spending

2026-01-20 · 34 mentions

#463 Issue #463: All about The

2026-02-02 · 20 mentions