The Archive

Every edition. Every call.

Daily briefings on AI and compute, published 06:00 UTC · each one closes with a falsifiable call · the record stands

The Book 151 editions 0 correct·0 wrong·0 push

16ThursdayJuly 2026 Latest edition The largest open model yet says it's Claude Moonshot's Kimi K3 is a 2.8-trillion-parameter open model that ships today claiming to beat Opus 4.8 at half the cost per finished task. Say hello and it says it's Claude, the tell of a distillation Anthropic put a number on in February. 9 sources · AI Call · Open Wed, Jul 15 The buildout's biggest balance sheet is one notch above junk 10 sources · COMPUTE · Call · Open Tue, Jul 14 "LOL, I found out I can access the [network storage], so funny." 9 sources · AI · Call · Open Mon, Jul 13 Nadella spent his Sunday telling enterprises they pay for AI twice 9 sources · AI · Call · Open Sun, Jul 12 Altman and Musk spent the weekend accusing each other of misleading public-market investors 9 sources · COMPUTE · Call · Open Sat, Jul 11 The frontier model finished turning into a commodity this week 9 sources · AI · Call · Open Fri, Jul 10 The biggest foreign listing in the history of the US stock market this week was not a chip designer or an AI lab. 9 sources · COMPUTE · Call · Open

July 2026

16 editions

MoTuWeThFrSaSu 1 Anthropic's cheap new model runs agents like the flagship did, and the price of capable AI fell again 9 src · AI 2 A Chinese open model showed up inside GitHub Copilot this week, one click from OpenAI 8 src · AI 3 "Right to Local Intelligence" got a manifesto this week, and the models to make it real 7 src · AI 4 On Independence Day, the story is a benchmark: 5 src · COMPUTE 5 Capable AI got cheaper all week. 9 src · AI 6 Meta's expensive GPUs were sitting idle, waiting on storage. 9 src · COMPUTE 7 Memory just repriced the AI build-out. 9 src · COMPUTE 8 A hidden line in a public GitHub issue turned the platform's own AI agent into a private-repo leak, and the industry spent… 9 src · AI 9 OpenAI's strongest model went public today after a 12-day government gate, and Grok matched its tier by afternoon at a… 9 src · AI 10 The biggest foreign listing in the history of the US stock market this week was not a chip designer or an AI lab. 9 src · COMPUTE 11 The frontier model finished turning into a commodity this week 9 src · AI 12 Altman and Musk spent the weekend accusing each other of misleading public-market investors 9 src · COMPUTE 13 Nadella spent his Sunday telling enterprises they pay for AI twice 9 src · AI 14 "LOL, I found out I can access the [network storage], so funny." 9 src · AI 15 The buildout's biggest balance sheet is one notch above junk 10 src · COMPUTE 16 The largest open model yet says it's Claude 9 src · AI 171819202122232425262728293031

June 2026

30 editions

MoTuWeThFrSaSu 1 OpenRouter Raises $113M Series B 37 src · DEV 2 Malicious npm Packages Found Inside Red Hat's Own Cloud Services Repos 30 src · DEV 3 Headroom: Cut 60-95% of Your LLM Token Costs With a Compression Layer 31 src · DEV 4 HeyGen's Hyperframes: Write HTML, Render Video 31 src · DEV 5 VoidZero Joins Cloudflare, The JS Toolchain Just Got a New Owner 28 src · DEV 6 Anthropic Open-Sources Its Vulnerability Discovery Framework 25 src · DEV 7 MemPalace: Open-Source AI Memory System Tops Benchmarks, Ships Free 20 src · DEV 8 OpenAI Codex merges into ChatGPT as revenue surges 50% weekly 35 src · DEV 9 Google Drops Agent Skills Framework 60 src · DEV 10 Anthropic splits the frontier: Fable 5 goes public, Mythos 5 goes behind a background check 60 src · DEV 11 DiffusionGemma breaks the token-by-token bottleneck: 56 src · DEV 12 Fable 5 tops every coding benchmark 60 src · DEV 13 Washington pulls two Anthropic models offline by federal order 32 src · DEV 14 Washington Orders Anthropic to Pull Fable 5 and Mythos 5, So It Pulled Them for Everyone 43 src · DEV 15 Fabro turns agent orchestration into a version-controlled graph with per-node model routing 34 src · DEV 16 Hetzner raises prices a third time in 2026, and AI is the reason 20 src · DEV 17 SpaceX buys Cursor for $60B and pulls developer tools inside the rocket 24 src · DEV 18 GLM-5.2 takes the open-weights crown at 51, and pays for it in tokens 34 src · DEV 19 Z.ai shipped GLM-5 under MIT and is already two releases past it 34 src · DEV 20 Builder.io ships an MIT-licensed framework that makes the agent a first-class user of your app 45 src · DEV 21 GLM-5.2 puts top-tier coding within four points of Claude for a sixth the cost 26 src · DEV 22 GLM-5.2 turns leaving Claude into a five-minute config change 39 src · DEV 23 SpaceX becomes a compute landlord, and Reflection signs a $6.3B lease 47 src · DEV 24 Claude burned through 2h37m of degraded models in a single day, and it is the third straight week 60 src · DEV 25 Anthropic put a remembering Claude inside Slack, and Salesforce owns the channel it learns on 7 src · AI 26 OpenAI launched its strongest model to about twenty government-approved partners, and no one else yet 9 src · AI 27 It was supposed to take six months. 9 src · AI 28 Wall Street goes hunting for the next Nvidia and lands on the people who make memory 24 src · COMPUTE 29 A Chinese open model caught Claude this week, and the moat that's left isn't the model 9 src · AI 30 Fable 5 is back at full capacity, and Anthropic's answer to a brutal week is a model you can actually call 8 src · AI

May 2026

26 editions

MoTuWeThFrSaSu12345 6 Bun Is Being Rewritten from Zig to Rust 32 src · DEV 7 Cloudflare Lets Agents Create Accounts, Buy Domains, and Deploy 29 src · DEV 8 Simon Willison: Vibe Coding and Agentic Engineering Are Converging Fast 24 src · DEV 9 AI Agents Get Structure: CloakBrowser, OpenSpec, and the Push for Agent Control Flow 32 src · DEV 10 ByteDance Ships Persistent Memory for AI Coding Agents 29 src · DEV 11 Bun's Rust Rewrite Hits 99.8% Test Compatibility 28 src · DEV 12 react-doctor: Lint Layer for AI-Generated React Code Hits 1.7K Stars Overnight 30 src · DEV 13 TanStack NPM Supply-Chain Compromise: 25 src · DEV 14 GitHub Launches Spec-Kit: Spec-Driven Development Gets a First-Class Toolkit 36 src · DEV 15 "Why Is This Running?", witr Explodes on GitHub as the Debugger Devs Actually Needed 23 src · DEV 16 Bun Rewrites Its Core in Rust, The Zig Era Is Over 35 src · DEV 17 CLI-Anything wants to make every piece of software agent-native via CLI wrappers 32 src · DEV 18 Agent Skills Registry Brings Package-Manager Discipline to AI Coding Agents 36 src · DEV 19 12-Factor Agents: A Production Manifesto for LLM-Powered Software 29 src · DEV 20 Karpathy Joins Anthropic, What It Signals for the Claude Ecosystem 28 src · DEV 21 Google I/O Model Blitz: Gemini 3.5 Flash Drops Alongside a Reinvented Search Box 35 src · DEV 22 3,800 GitHub Repos Breached via Malicious VSCode Extension 32 src · DEV 23 Anna's Archive Plants an llms.txt That Talks Directly to AI Crawlers 26 src · DEV 24 Anthropic's Project Glasswing: What We Know and What It Means for Builders 26 src · DEV 25 Anthropic Open-Sources Claude Cowork Plugins 29 src · DEV 26 Uber COO: "Harder to Justify" AI Token Spend 26 src · DEV 27 Microsoft Ships Agent Governance Toolkit 30 src · DEV 28 stop-slop: A Skill File That Strips AI Tells From Your Prose 28 src · DEV 29 Claude Opus 4.8 Drops Alongside Dynamic Workflows in Claude Code 26 src · DEV 30 3,000 tokens/s on standard GPUs: 29 src · DEV 31 SQLite Is All You Need for Durable Workflows 25 src · DEV

April 2026

29 editions

MoTuWeThFrSaSu 1 Axios NPM Package Compromised, Malicious Versions Drop Remote Access Trojan 26 src · DEV 2 Claude Code Source Leak Reveals Fake Tools, Frustration Detection, and Undercover Mode 28 src · DEV 3 Google Drops Gemma 4 Open Models + Qwen 3.6-Plus Ships Agent-First Architecture 41 src · DEV 4 Cursor 3 Drops: The AI Code Editor Gets Its Biggest Overhaul Yet 28 src · DEV 5 Mintlify Ditched RAG for a Virtual Filesystem 34 src · DEV 6 Codex Moves to API-Based Pricing 27 src · DEV 7 Gemma 4 Goes Local: On-Device LLMs Hit iPhone and Desktop via LiteRT-LM 29 src · DEV 8 Goose: Open-Source AI Agent That Actually Ships Code, Not Just Suggests It 33 src · DEV 9 Anthropic Launches Project Glasswing: 33 src · DEV 10 Daytona Ships Secure Sandboxing for AI-Generated Code at Scale 30 src · DEV 11 InstantDB 1.0: A Backend Built for the AI-Coded App Era 31 src · DEV 12 Google Open-Sources MCP Toolbox for Databases 22 src · DEV 13 Ralph: The Autonomous Agent Loop That Ships Your PRD While You Sleep 28 src · DEV 14 Open Generative AI: Self-Hosted Image & Video Studio With 20+ Models, MIT Licensed 25 src · DEV 15 GitHub Ships Stacked PRs, Finally Catching Up to How Teams Actually Ship Code 21 src · DEV 16 Vercel Ships Open-Source Cloud Agent Template 34 src · DEV 17 Claude Opus 4.7 drops, Anthropic's biggest model leap in a year 32 src · DEV 18 Google Ships Android CLI, Any AI Agent Can Now Build Android Apps 3x Faster 28 src · DEV 19 Thunderbird Launches Thunderbolt: 30 src · DEV 20 Zero-Copy GPU Inference from WebAssembly on Apple Silicon Changes the Edge AI Game 34 src · DEV 21 Manifest: Smart Model Routing Cuts Agent Costs Up to 70% 38 src · DEV 22 Claude Code Gets Full-Codebase Context via Zilliz MCP 32 src · DEV 23 Qwen3.6-27B Drops: Flagship-Level Coding in a Model You Can Actually Self-Host 33 src · DEV 24 Context-Mode Cuts 98% of AI Agent Context Bloat Across 12 Platforms 31 src · DEV 25 GPT-5.5 Drops While DeepSeek v4 Launches 30 src · DEV 26 Matt Pocock's .claude Skills Directory: 27 src · DEV 27 OpenAI Retires SWE-bench Verified 30 src · DEV 28 Microsoft and OpenAI End Exclusivity 35 src · DEV 29 GitHub Copilot Code Review Now Eats Your Actions Minutes 30 src · DEV 30

March 2026

31 editions

MoTuWeThFrSaSu 1 Agent Orchestration Goes Multi-Player: 32 src · DEV 2 Qwen3.5 Hits Sonnet 4.5 Performance at 35B 40 src · DEV 3 WebMCP Brings Model Context Protocol to the Browser 38 src · DEV 4 WiFi-Based Pose Estimation Goes Open Source 30 src · DEV 5 Shannon: Autonomous AI Hacker Hits 96% Exploit Success Rate on XBOW Benchmark 37 src · DEV 6 Nvidia PersonaPlex 7B Runs Full-Duplex Speech-to-Speech on Apple Silicon 34 src · DEV 7 GPT-5.4 Drops, OpenAI's Incremental Naming Hides a Capability Jump 36 src · DEV 8 Anthropic Red-Teams Firefox: AI-Powered Security Auditing Goes Mainstream 34 src · DEV 9 Run Qwen 3.5 Locally with Unsloth 29 src · DEV 10 Agent Safehouse: macOS-Native Sandboxing for Local AI Agents Is Here 32 src · DEV 11 Docker Ships an Official AI Agent Builder and Runtime 32 src · DEV 12 Cloudflare Ships a Crawl Endpoint 36 src · DEV 13 Agent Memory Gets Real: Three New OSS Projects Attack RAG From Different Angles 30 src · DEV 14 Lightpanda: A Headless Browser Built Ground-Up for AI Agents and Automation 32 src · DEV 15 Claude Opus 4.6 and Sonnet 4.6 Ship 1M Context Window to GA 33 src · DEV 16 Claude Code Best Practices Repo Explodes 31 src · DEV 17 LangChain Ships DeepAgents: Planning, Subagents, and Filesystem Access Out of the Box 29 src · DEV 18 Mistral Drops Leanstral: An Agent That Writes Formal Proofs Alongside Code 32 src · DEV 19 Everything Claude Code: The Agent Harness Framework Dominating GitHub This Week 40 src · DEV 20 Astral Joins OpenAI, Python's Best Tooling Team Gets Absorbed 28 src · DEV 21 OpenCode and oh-my-openagent: Open-source coding agents are having their moment 36 src · DEV 22 Open Source AI Coding Agents Hit Critical Mass: 26 src · DEV 23 Flash-MoE Runs a 397B Parameter Model on a Laptop 40 src · DEV 24 Pascal Editor: Open-Source Autonomous Agents Hit 11K Stars Overnight 33 src · DEV 25 LiteLLM Supply-Chain Attack: If You Installed It Recently, Check Now 34 src · DEV 26 Sora Is Dead. OpenAI's Video Generator Shuts Down. 30 src · DEV 27 oh-my-claudecode: Multi-Agent Orchestration Lands for Claude Code Teams 33 src · DEV 28 Microsoft Open-Sources VibeVoice: 35 src · DEV 29 Cursor Ships Real-Time RL for Composer 32 src · DEV 30 Claude Code Gets Its Cookbook: 5,600+ Stars on a Copy-Paste Agent Guide 29 src · DEV 31 MCPorter: Call Any MCP Server as a Plain TypeScript Function or CLI 32 src · DEV

February 2026

19 editions

MoTuWeThFrSaSu123456789 10 GPT-5.3-Codex Launches in GitHub Copilot and Cursor 60 src · AI 11 Stripe's AI 'Minions' Now Merge 1,000+ Pull Requests Per Week 60 src · AI 12 GLM-5 Drops: 744B Open-Weight Model Under MIT License Shakes Up the Coding Leaderboard 60 src · AI 13 Google's Gemini 3 Deep Think Hits 84.6% on ARC-AGI-2, Rewriting the Reasoning Ceiling 60 src · AI 14 Open-Source AI Models Storm the Leaderboards as Free Agentic Coding Goes Mainstream 32 src · AI 15 DeepSeek v4 Drops Next Week, Open-Source May Finally Match Frontier 51 src · DEV 16 OpenClaw hits 12K+ stars: the open-source personal AI assistant wave is real 48 src · DEV 17 Qwen3.5 drops as a native multimodal agent model 52 src · DEV 18 Claude Sonnet 4.6 drops with 1M context, 15-point perf jump, and GitHub Copilot integration 60 src · DEV 19 Claude Code Now Pushes Prototypes Directly to Figma via MCP 52 src · DEV 20 Gemini 3.1 Pro drops: 77% ARC-AGI-2, matches Opus 4.6 on SWE-Bench 60 src · DEV 21 Gemini 3.1 Pro drops, Google's biggest model update lands with 2K+ HN points 53 src · DEV 22 Karpathy Introduces "Claws", A New Mental Model for AI Agent Capabilities 47 src · DEV 23 Leaked System Prompts for 30+ AI Coding Tools Now in One Repo 36 src · DEV 24 PageIndex: Persistent Memory for Always-On AI Agents Hits 2.7K Stars 34 src · DEV 25 Every Major AI Coding Tool's System Prompt Just Leaked 40 src · DEV 26 Claude Code Gets Remote Control 40 src · DEV 27 SpacetimeDB: The Database That Wants to Kill Your Backend 32 src · DEV 28 Anthropic vs. the Department of War: 34 src · DEV