-
OpenMontage hits 53,240 GitHub stars: your coding agent is now a video studio
OpenMontage is not an app. It’s 700+ agent skill files, 12 YAML production pipelines, and 100+ tools that drop into Claude Code, Cursor, or Copilot — and turn the coding agent into a full video production crew. Describe the video in plain English; the agent researches the topic, writes the script, generates assets, edits the… Continue reading
-
archify adds 4,561 GitHub stars in one day — the agent skill that stops your coding agent from hand-drawing diagrams
archify is a Node.js agent skill for Cursor, Claude Code, Codex CLI, and OpenCode. One npx skills add tt-a1i/archify -g and your coding agent turns a codebase or a plain-text description into interactive system diagrams — architecture, workflow, sequence, data flow, lifecycle — shipped as a single self-contained HTML file. 27,093 stars, GitHub Trending #1… Continue reading
-
Hugging Face launches Microduck, a $399 open-source bipedal robot you train yourself
Hugging Face’s Pollen Robotics opened pre-orders for Microduck: a 25cm, sub-800g bipedal robot with 15 motors, a camera, LiDAR, two IMUs, and an articulated beak that lifts 800g. It waddles, crouches, roller-skates, and stands back up after falls. $399, shipping before Christmas to North America, Europe, and the UK. Why $399 is the real story… Continue reading
-
US judge rules Pentagon blacklisting of Anthropic unlawful — the first courtroom win for an AI safety red line
Anthropic refused to strip Claude’s guardrails against fully autonomous weapons and mass surveillance of Americans. The Pentagon’s answer in February: designate the company a “supply chain risk” — the first American company ever hit with that label — and bar every defense contractor from doing business with it. This week, federal judge Rita F. Lin… Continue reading
-
Spline V2 rewrites the entire 3D editor — then hands the keys to Claude Code and Cursor
Spline shipped V2 on August 21: a full rebuild of its 3D design editor. New UI, WebGPU renderer with WebGL fallback, PBR/HDR pipelines, boolean operations 4x faster, plus native macOS and Windows apps. The rebuild is the footnote. The headline is who gets to drive it. An editor built for agents, not just designers V2… Continue reading
-
Claude Cowork Built-In Browser (Anthropic) lands two weeks after OpenAI killed Atlas
Anthropic just gave Claude its own browser. Starting August 26, when a task in the Cowork desktop app touches the web, a browser opens in the side panel and Claude navigates, reads, clicks, and types on its own — filling forms, pulling numbers from dashboards, working through portals that have no connector. What it actually… Continue reading
-
Gemini Omni 1.1 Flash hits GA: Google prices 4K AI video at $0.30 a second
Google’s video generation and editing model went GA on August 27, and the pricing is the real story: $0.03 per second for 360p, $0.10 for 720p, $0.15 for 1080p, $0.30 for 4K. A 360p draft mode runs 60% faster than 720p at a third of the cost — iterate cheap, upscale once. What’s new in… Continue reading
-
Gemini 3.5 Transcribe (Google) hits 2.6% word error rate — and deletes your “ums” for you
Apple’s SpeechAnalyzer just beat Whisper a few months ago. Now Google fires back. Gemini 3.5 Transcribe launched August 26 and immediately hit both Product Hunt and the HackerNews front page (181 points). It’s a dedicated speech-to-text model, not a chatbot feature: 2.6% word error rate on recorded audio, 4.0% streaming, 70% lower latency than its… Continue reading
-
GLM-5.3-Flash (Z.ai) is Ox Alpha — the mystery model that hit 2 on OpenRouter running on Chinese chips
Mystery solved. Ox Alpha, the anonymous model that processed 16 trillion tokens for 221,000 users in one week on OpenRouter, is GLM-5.3-Flash from Z.ai. The kicker: the whole thing ran on domestic Chinese AI accelerators. The stealth launch was a production stress test, and it passed in public. What it actually is The first natively… Continue reading
-
Anthropic Model Hardware Standard (MHS) compresses weeks of lab hardware integration into hours
Anthropic opened the first research preview of the Anthropic Model Hardware Standard (MHS) on August 27 — a protocol that lets AI agents safely operate physical equipment: microscopes, liquid handlers, robotic arms, anything with a programmable interface. Think USB-C, but between agents and machines. The pain it kills: wiring an agent to lab hardware today… Continue reading
