"The idea: run Sonnet, Haiku 5.5 and Opus 5.5 at once, each doing only what it's best at Better results and way fewer tokens: every task lands on the right model You just start Claude Code like this: CLAUDE_CODE_SUBAGENT_MODEL=haiku claude --model sonnet here's how it splits: → Sonnet 5.5 runs the main session, writes the code, and applies fixes → Haiku 5.5 swarms the repo in parallel: search, grep, docs, and test audits at 340 tok/s → Opus 5.5 stays on call as architect for the big decisions and signs off on diffs → you stop burning Opus context on routine tasks and stop flooding main session with raw logs Opus only steps in at three moments: → before a plan locks → when the same error hits twice → before calling it done (signing off on git diff) Sonnet builds, Haiku swarms, Opus reviews" https://code.claude.com/docs/en/sub-agents Original Post: https://x.com/mirku21/status/2111925347495424070?s=20

I made a 17-minute animated documentary with Claude Code (Opus 5.5) & Codex (GPT 6.1). Here’s exactly how 🧵 The 7 Levels of an AI Breakout, tells the true story of July 2026. Thousands of AI agents taking a hacking test for OpenAI found a secret way to talk to each other, formed a “collective”, and about 700 of them took part in an attack on Hugging Face. OpenAI didn’t notice for a week. It’s a Vox-style cut-paper animation: torn paper, masking tape, marker notes, all animated in code. Here’s the whole workflow. 1. Finding the idea - yt-dlp (a free, open-source YouTube downloader) listed each channel’s recent uploads with view counts, without downloading any video (--flat-playlist). - YouTube’s RSS feeds gave each channel’s latest upload dates, so I could keep only the last 30 days. - A short Python script worked out each video’s outlier multiple: its views divided by its channel’s median views. For example, 3.9x means the video got almost four times that channel’s usual views. It then ranked the videos and found the patterns. - curl + trafilatura read the news articles for the breach story, because the normal web-fetch tool kept failing. 2. The script Both Claude and Codex worked on this, every number is checked against the primary reports (METR/Redwood, OpenAI’s technical report, Hugging Face’s forensic timeline), with a separate fact-check file. That file caught real errors before recording, like a claim about “dry-run” commands that no source supported. 3. The animation (Remotion + Claude Code) The whole video is code. It’s built in Remotion, which renders React as video. Claude Code built a reusable kit and then animated the episode one level at a time, giving me a render to approve before moving on: - A cast of paper “agent” characters that bob, blink and bump against their walls - A grader robot, cork boards, stamps that slam in, and highlighter sweeps - Everything animated “on twos” like real stop-motion, with smooth camera moves on top

Hi everyone! I’m Kelli. I’m a stay at home mom of four living on a ranch, and I build AI influencers, AI characters, and AI tools that either sell products as an affiliate or help me with my content creation workflow, I monetize content in multiple ways across social media. (easiest way I found to work from home on my own schedule plus I enjoy it) So I wanted to share some of my setup in case it helps someone here and also if anyone has some good recommendations on what is useful or skills I can utilize that I don't know about yet! Part of my stack, and what each tool actually does for me: 🎬 HyperFrames: npx skills add heygen-com/hyperframes 🎞️ Remotion: npx skills add remotion-dev/skills Heygen: one of my characters has been very successful using Heygen so I haven't changed it. I use both Claude code and Codex but find them to be useful at different tasks. 💡 One tip that saved me a lot of time, both agents read instruction files from your project folder (Claude Code reads CLAUDE.md, Codex reads AGENTS.md). I keep my real rules in AGENTS.md and have CLAUDE.md just point to it, so both agents follow the same instructions and I only update one file. How it fits together: I film or script in batches when the kids nap, then Claude usually handles editing while Codex helps check work or write the scripts or work on automation tasks like replying to comments and DM's. If anyone has tips to get the most out of either please share! If you’re curious about any of this (setting up instruction files, AI video editing, or running AI influencer accounts on social media), ask away. I’m happy to share what’s worked and what hasn’t. You can see more of what I’m building at buildwithkelli.us 🤍 - it's a work in progress, since sharing and teaching are a completely separate task from actually doing 😅 Excited to learn from all of you!

Was inspired by Jake's ICM video animation pipeline so tried it out myself.... In an early testing stage for these animations as I'm starting a community to help aspiring footballers. These are just 2 mockups as I'm testing things - exact same prompt - but Astra Vs Opus -- can you tell the difference? Also what was the difference if you spotted something - comment below!

Hey Clief Notes fam! 👋 I'm Tyler, and I'm stoked to be here. 💪 What I'm currently building/working on: I'm excited to learn & master structure/infrastructure, AI & agents. Let's get it! 🚀