Andrej Karpathy Just Revealed His New AI Workflow… And It’s Not What You Think 🤯

You are helping me turn one specific YouTube video into real, durable learning.

Source video:
- Title: Andrej Karpathy Just Revealed His New AI Workflow… And It’s Not What You Think 🤯
- URL: https://www.youtube.com/watch?v=eMVaGOsfYs8
- Topic: Interfaces + Open Design
- My current learning frame: Create a bot in your team's messaging app, wire it to a small server that forwards messages to an AI agent loaded with your business context, connect one real tool (calendar or email) through an integration layer like Compozio, and have two teammates tag it in the same thread to complete a shared task.
- Why this matters: New playlist item from Dream Labs AI; queued for transcript-backed review, topic mapping, and a practical learning artifact.

Transcript anchors from this exact video:
- 0:15 / Evidence 1: "not a new prompt, it's not a slightly better chatbot. It is, in his words, a completely different way of working. He described this third AI shift on Twitter as a system that is way more integrated with..."
- 2:08 / Evidence 2: "you talk to Claude directly within the browser. The second big redesign was that it's an app that you download. You download Claude code to your computer. It has access to your computer and all your files and..."
- 3:39 / Evidence 3: "where Andre Kapathy sees this going over the next couple of months. Because it is only available to the enterprise level, a lot of people are going to be steering clear of this, but we all want to..."
- 5:14 / Evidence 4: "own browser, uses their Claude bot to go edit it from there. That was evolution one. Evolution two, however, was a big breakthrough. This is when you could actually download Claude or Codex or whatever AI you use..."
- 7:27 / Evidence 5: "says, you need to do all the under the hood engineering work to make this just work. Get it across tools, get it across integrations, get it across compute environments, memory and security. And the problem is it's..."
- 10:06 / Evidence 6: "Claude. For the big jobs, it spins up this agent and then we get a fresh AI worker with the business brain, with the business context, with all the back business data, plugged into the right tools because..."

Your task:
1. Use the transcript anchors above as the primary source packet. If you add outside context, label it clearly as outside context and keep it secondary.
2. Create a source-check table with columns: timestamp, claim, what the demo proves, confidence, and what still needs verification.
3. Extract the actual teachable claims from the video. Do not invent claims that are not supported by the title, lesson frame, or transcript anchors.
4. Build a reusable learning artifact: A UI critique sheet for judging whether an AI interface improves control.
5. Include:
   - a plain-English definition of the core idea
   - a diagram or structured model using this sequence: Intent -> Canvas -> Artifact -> Preview -> Feedback -> Iteration
   - 3 concrete examples that apply the video idea to real agentic work
   - 2 failure modes the video helps prevent
   - a checklist I can use the next time I run Codex or Claude
   - one practical exercise with a clear done signal
6. Add a "learning transfer" section: what changes in my workflow tomorrow if I actually learned this?
7. Add a "source check" section that cites which transcript anchor supports each major takeaway.

Quality bar:
- Make this specific to "Andrej Karpathy Just Revealed His New AI Workflow… And It’s Not What You Think 🤯", not a generic Interfaces + Open Design essay.
- Prefer operational examples, failure modes, and reusable artifacts over broad definitions.
- Call out uncertainty instead of smoothing over weak evidence.
- If evidence is weak, say what transcript segment or timestamp needs review instead of guessing.
- Finish with a concise artifact I could paste into my learning app.