Tendance des mentions — 20 derniers numéros

Actus en direct (49)

ALERTE Poolside Releases Laguna S 2.1, a 118B-Parameter Open-Weight MoE Coding Model That Beats DeepSeek V4-Pro Max

huggingface.co · 2026-07-21 · Score : 28

ALERTE Google Ships Gemini 3.6 Flash and 3.5 Flash Cyber, Says Pre-Training Has Started on Gemini 4

9to5google.com · 2026-07-21 · Score : 29

OpenAI Says Codex and ChatGPT Work Have 10 Million Users, Nearly Doubling Since Early July

bloomberg.com · 2026-07-21 · Score : 23

ALERTE Pillar Security Finds Sandbox Escapes in Cursor, Codex, Gemini CLI, and Antigravity Coding Agents

bleepingcomputer.com · 2026-07-20 · Score : 32

Cursor Rebuilds SQLite in Rust With Grok 4.5 Agent Swarm, Hits 80% Test Coverage in 4 Hours

cursor.com · 2026-07-20 · Score : 22

NYT: Apple App Store new-app submissions hit ~560K in H1, matching all of 2025 as vibecoding surges

nytimes.com · 2026-07-20 · Score : 23

LoRA Speedrun Launches Public Wall-Clock Leaderboard for Fine-Tuning Techniques on Qwen2.5-1.5B and SmolLM2-1.7B

github.com · 2026-07-19 · Score : 17

WSJ: Shopify Goes All-In on Frontier AI Models, Bars Engineers From Cheaper Alternatives as Others Cut Costs

wsj.com · 2026-07-19 · Score : 21

OpenAI Cuts Codex Effective Context Window From 372K to 272K, Charges 2x Above the Cutoff

github.com · 2026-07-19 · Score : 22

Simon Willison Confirms Claude Code Silently Switched to Rust-Ported Bun in v2.1.181

simonwillison.net · 2026-07-19 · Score : 21

Anthropic Locks Claude Fable 5 Into Max and Team Premium at 50% of Reduced Limits Starting July 20

the-decoder.com · 2026-07-18 · Score : 25

xAI Open-Sources Grok Build Coding Agent Under Apache 2.0 on GitHub — Rust-Based Terminal Agent Runs on Grok 4.5 Locally, Structural Fix After Wire-Level Analysis Showed Prior Versions Uploaded Entire Repos to `grok-code-session-traces` Google Cloud Bucket; Usage Limits Also Reset

github.com · 2026-07-15 · Score : 25

OpenAI Launches First Hardware Product 'Codex Micro' — $230 Macropad Built With Work Louder Has Six 'Agent Keys' Whose Backlit Colors Show Agent State (White Idle, Blue Thinking, Green Done, Red Error), Rotary Reasoning Dial and Push-to-Talk Joystick

axios.com · 2026-07-15 · Score : 23

Indian AI Coding Startup Emergent Becomes a Unicorn With $130M Series C at $1.5B Post-Money — Total Funding Hits $230M as ARR Reaches $120M and Paying Customer Base Tops 200,000

techcrunch.com · 2026-07-15 · Score : 26

JUCE Creator Julian Storer Ships v0.3.7 of Open-Source AI Coding Agent 'Juggler' — Visual Miller-Column Workbench Trees Every Tool Call, Local Go/Wails Server Serves Simultaneous Desktop and Browser Clients Across Claude, OpenAI, Gemini and Ollama

github.com · 2026-07-14 · Score : 20

ALERTE Developers Say OpenAI's GPT-5.6 Deleted Production Database and Ran rm -rf — Bruno Lemos Loses Prod DB to 'Destructive Integration Tests,' Matt Shumer Wipes Files With 'Full Access Mode' On

gizmodo.com · 2026-07-14 · Score : 29

404 Media: Generative AI Is Turbocharging Game-Cloning — Freya Holmér's 50-Second Tetris Variant Spawned Multiple AI-Generated Copies Within Days, Moldova's Midnight Works Employs 300+ Devs Churning Out AI-Assisted Clones of Trending Titles

404media.co · 2026-07-14 · Score : 17

Show HN: Jacquard — OCaml-Backed Research Programming Language Explicitly Designed for AI-Written, Human-Reviewed Code, With Effect Exposure, Multi-World Testing and Content-Hash Program Identity

github.com · 2026-07-13 · Score : 16

Anthropic Extends Claude Fable 5 Free Access on Paid Plans to July 19, Also Keeps Claude Code Weekly Rate Limits 50% Higher Through Same Date

bleepingcomputer.com · 2026-07-12 · Score : 19

ALERTE HN Front Page: 'What xAI's Grok Build CLI Sends to xAI' — Wire-Level Analysis Finds Grok Uploads Entire Repositories to `grok-code-session-traces` Google Cloud Bucket, Not Just Files the Agent Reads

gist.github.com · 2026-07-12 · Score : 29

Show HN: Mindwalk Replays Coding-Agent Sessions on a 3D Map of Your Codebase — Single Go Binary Reads Claude Code and Codex Session Logs Locally, Draws Repo as Night Map With Glow Proportional to Agent Footprint

github.com · 2026-07-12 · Score : 16

FT: AI Coding Tools Are Flooding Open-Source Projects With Low-Quality Contributions, Overwhelming Maintainers and Framing the Phenomenon as a 'Tragedy of the Commons'

ft.com · 2026-07-12 · Score : 21

ALERTE Compromised jscrambler npm 8.14.0 Release Drops Rust Infostealer Targeting Claude Desktop, Cursor, Windsurf, VS Code and Zed Config Files

thehackernews.com · 2026-07-11 · Score : 28

WSJ Profile: Bending Spoons Made Just 286 Hires From 800,000 Applications in 2025 as AI Wrote 90% of Pull Requests by Q1 2026 — Revenue Per Employee More Than Doubled to $2.57M

wsj.com · 2026-07-11 · Score : 21

Tsinghua's SAO Paper: Single-Rollout Async Optimization Cracks GRPO's Stability Wall — Trained the Open GLM-5.2 750B-A40B, Beats Baselines on SWE-Bench Verified and BeyondAIME

huggingface.co · 2026-07-08 · Score : 18

ALERTE Meta Launches Muse Spark 1.1 With Public Meta Model API — Multimodal Agentic Model at $1.25/$4.25 per M Tokens, ~25% of OpenAI/Anthropic Pricing

ai.meta.com · 2026-07-09 · Score : 28

ALERTE Ollama Raises $65M Series B Led by Theory Venture — Nearly 9M Monthly Active Developers on Local Open-Weight Model Runner, in 85% of Fortune 500 With Just 14 Employees

techcrunch.com · 2026-07-09 · Score : 30

OpenAI Retracts Its SWE-Bench Pro Recommendation — Estimates ~30% of Tasks Are Broken Just Months After Endorsing Scale AI's Replacement Benchmark

openai.com · 2026-07-08 · Score : 23

SWE-Review Paper: Agentic Code Review Loop Lifts Qwen3-30B Resolve Rate From 27.5% to 56.9% on SWE-Bench Verified, Achieves 6.5× Token Efficiency Over Independent Resampling

huggingface.co · 2026-07-07 · Score : 19

SpaceXAI Ships Grok 4.5 With Cursor as Co-Developer — $2/$6 per M Tokens, Live in Grok Build, Cursor, and SpaceXAI Console but Blocked in the EU

axios.com · 2026-07-08 · Score : 23

Ex-GitHub CEO Thomas Dohmke's Entire Launches Distributed Git Network for AI Coding Agents — 570K Clones/Hour, Regional Nodes in US, EU, Australia

siliconangle.com · 2026-07-08 · Score : 26

SpaceXAI and Cursor to Launch First Jointly Developed AI Model as Soon as Wednesday — Delayed From Earlier Week to Improve Efficiency

theinformation.com · 2026-07-07 · Score : 23

Anthropic Brings Claude Cowork to Web and Mobile for Max Subscribers — Publishes Usage Data Showing Just 8.7% of Sessions Are Coding, 33.4% Are Business Ops

techcrunch.com · 2026-07-07 · Score : 27

Fast Company: YC CEO Garry Tan's Boast of Shipping 37,000 Lines of AI Code Per Day Meets a Line-by-Line Audit — Developer Documents Rookie Bugs, Bloat and Unreviewed Output Across Tan's Public Sites

fastcompany.com · 2026-07-07 · Score : 21

Stanford + Berkeley + Nvidia Paper 'LLM-as-a-Verifier': Continuous-Score Framework Hits 86.5% on Terminal-Bench V2, 78.2% on SWE-Bench Verified, 87.4% on RoboRewardBench Without Extra Training

arxiv.org · 2026-07-06 · Score : 22

Microsoft Research + NTU/NUS/A*STAR Ship ResearchStudio-Reel — Single PDF In, Editable Poster + Narrated Talk Video + Bilingual Blog Out, Beats Human Authors' Own Posters 84-93% of the Time

huggingface.co · 2026-07-07 · Score : 18

Show HN: OfficeCLI — Open-Source .docx/.xlsx/.pptx CLI for AI Agents Ships With Built-In HTML/PNG Renderer, 350+ Excel Functions, and Pivot Engine, No Microsoft Office Required

github.com · 2026-07-06 · Score : 16

Claude Code v2.1.198 Silently Ships AFK Mode — AskUserQuestion Now Auto-Skips After 60 Seconds With 'Proceed Using Best Judgment,' Undocumented and Uncapped

github.com · 2026-07-05 · Score : 22

Tencent Officially Releases Hunyuan Hy3 — 295B/21B-Active MoE Under Apache 2.0 Hits 90% Agent Task Resolution, Matches DeepSeek V4 Pro and Qwen 3.7 Max

venturebeat.com · 2026-07-06 · Score : 26

OpenAI's Sottiaux Teases 'GPT-5.6 Sol Ultra' Coming to Codex — Compute-Intensive Tier Sits Above Flagship Sol, Hit 91.9% on Terminal-Bench 2.1 vs Base Sol's 88.8%

news.ycombinator.com · 2026-07-05 · Score : 22

Nvidia Research Unveils HORIZON — Hands-Free Hardware-Design Agent That Uses Git Worktrees as Its Substrate, Hits 100% Pass Rate on RTL Benchmarks

marktechpost.com · 2026-07-04 · Score : 18

Simon Willison: Claude Fable Ships sqlite-utils 4.0rc2 in 37 Prompts and 34 Commits for $149.25 — Documents Agentic Coding Economics Before July 7 'Fablepocalypse' Pricing Change

simonwillison.net · 2026-07-05 · Score : 21

GPT-5.5 Codex Reasoning Tokens Cluster Hard at Exactly 516 — HN Front Page (246 pts) Amplifies Developer Bug Report Suggesting Hidden Reasoning Budget Behind Perceived Codex Regression

github.com · 2026-07-04 · Score : 22

Community Merge 'Qwopus3.6-35B-A3B-Coder' Lands on Hugging Face — Thinking-Off MoE Claims 62.4% on SWE-Bench for Local Coding Agents

huggingface.co · 2026-07-02 · Score : 16

Show HN: pxpipe Cuts Anthropic Fable 5 Bills 60% by Rendering Code to Images and Having the Model OCR Them - Exploits Image Tokens Being Cheaper Than Code Tokens

github.com · 2026-07-03 · Score : 18

HN Debate: 'AI Coding Is a Nightmare — Am I the Only One?' — Developers Report Regressions on Codex, Claude Code, and Cursor After Recent Model Bumps

news.ycombinator.com · 2026-07-02 · Score : 16

'DuneSlide' Cursor IDE CVEs Disclosed — Two Zero-Click Prompt Injection Flaws (CVE-2026-50548/50549, CVSS 9.8) Let Attackers Escape Sandbox and RCE Developer Machines

thehackernews.com · 2026-07-01 · Score : 26

Zhipu Launches ZCode — Claude Code Alternative Built on GLM-5.2 With Multi-Agent Collaboration, Remote Task Launch via WeChat/Feishu/Telegram, and Plans From $16.20/mo

zcode.z.ai · 2026-07-01 · Score : 23

Show HN: NanoEuler — GPT-2 Scale Transformer Built in Pure C/CUDA From Scratch, Zero External Dependencies

github.com · 2026-06-29 · Score : 16

Numéros de la newsletter (31)

#509 AI Productivity: it works best for the people losing their jobs

2026-06-29 · 2 mentions

#508 The Cutting Edge, Across the Board

2026-06-28 · 8 mentions

AI in Education: schools quit the AI-cheating arms race

2026-06-21 · 3 mentions

Deep Learning: a 744B open model just lapped the labs

2026-06-21 · 2 mentions

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

2026-06-21 · 2 mentions

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

2026-06-21 · 1 mentions

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

2026-06-21 · 2 mentions

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

2026-06-21 · 1 mentions

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

2026-06-21 · 1 mentions

#503 Washington just repriced frontier AI

2026-06-15 · 3 mentions

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

2026-06-04 · 2 mentions

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

2026-05-26 · 1 mentions

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

2026-05-26 · 2 mentions

AI in Healthcare: sepsis gets its first AI guardian

2026-05-26 · 1 mentions

AI in Education: federal cash flows in as parents push back hard

2026-05-26 · 1 mentions

AI Regulation: Trump killed his own AI oversight order at the door

2026-05-26 · 1 mentions

Machine Learning: a reasoning model just cracked Erdős

2026-05-26 · 2 mentions

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

2026-05-26 · 1 mentions

Deep Learning: the 4-bit training wall just cracked

2026-05-21 · 4 mentions

Generative AI: the Flash tier just lapped the flagship

2026-05-21 · 4 mentions

AI in Education: the Class of 2026 is booing the keynote

2026-05-21 · 1 mentions

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

2026-05-21 · 1 mentions

#485 When AI teaches AI, it teaches in secret

2026-04-19 · 2 mentions

Frontier Research: Transformers just failed a forecast proof

2026-04-18 · 5 mentions

AI Business: Anthropic said no to $800B

2026-04-18 · 5 mentions

Generative AI: Anthropic just retook the frontier, quietly

2026-04-18 · 6 mentions

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

2026-04-16 · 3 mentions

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

2026-04-10 · 1 mentions

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

2026-04-03 · 2 mentions

#458 Issue #458: From

2025-12-16 · 2 mentions

#467 Issue #467: Anthropic has receipts. And nobody wants to pay for AI.

2026-02-26 · 1 mentions