The New AI prompting Method Everyone Is Talking About: Loops (And How To Use Them)

Boris Cherny (the creator of Claude Code) and Peter Steinberger (creator of
OpenClaw) have both said this is now how they program. They don't prompt a
chatbot or write code by hand, they build loops where the AI prompts itself.

They're calling it the future. It's also super confusing if you're just
hearing about it. So this video breaks down exactly what loops are, how to
use them, what they're good for, what they're NOT good for, and walks through
real demos so you can start using this yourself.

You don't have to be technical to follow along. If a terminal window scares
you, don't worry, I do the whole thing inside the Claude Code desktop app.

We build a real-world example from scratch: growing an online pickleball
store using 3 AI agents running at the same time. One builds a Harry Potter
x pickleball personality quiz, one scouts content opportunities across Reddit
and competitor sites, and one turns it all into ready-to-use marketing
(emails, social captions, site recommendations). Then I show you how to hand
the whole thing off to an orchestrator agent so it runs on a weekly loop
automatically. We finish with a 4-scenario quiz where you figure out how
you'd loop each one.

One note: I made this just before Anthropic released Claude Fable 5
(June 9, 2026), and that launch makes everything in here matter even more.
Fable 5 is Anthropic's first Mythos-class model, built specifically for the
long-running, multi-stage autonomous tasks that loops depend on, like planning
across stages, delegating to sub-agents, and checking its own work for hours
at a time. The method I'm showing you here just got a serious upgrade.

🔑 WHAT YOU'LL LEARN

- The difference between prompting and looping
- How to run multiple AI agents in parallel in the Claude Code desktop app
- The core stages of every loop: discovery, planning, execution,
verification, iteration
- Open vs closed loops — and which one won't burn through your budget
- How to set up an orchestrator agent to manage a fleet on a weekly schedule
- 4 real-world scenarios you can loop today

🛠️ TOOLS MENTIONED
Claude Code · Claude desktop app · OpenClaw · Figma
(Note: you'll need a Claude Pro plan or higher to access Claude Code)

📌 KEY PEOPLE
Boris Cherny (Claude Code) · Peter Steinberger (OpenClaw)

This is the shift from holding the tool to designing the system. If you use
AI for work, content, coding, research, or running a business, this is the
mental model you need for the rest of 2026.

⏱️ TIMESTAMPS
0:00 - Why the top AI programmers stopped prompting
0:44 - What a loop actually is (the old way vs the new way)
1:28 - The 5 stages of a loop, explained visually
2:45 - Open loops vs closed loops (and which to use)
3:35 - Real-world example: growing a pickleball store
5:00 - The 3-agent system, mapped out visually
7:16 - Building it live in the Claude Code desktop app
7:28 - Agent 1: the quiz builder prompt
8:07 - Agent 2: the content scout prompt
9:21 - Agent 3: the growth marketing agent prompt
11:18 - Checking the results:
13:38 - How to automate it all with an orchestrator agent
15:35 - QUIZ: how would you loop these 4 scenarios?
15:47 - Scenario 1: the freelancer
16:39 - Scenario 2: the student
17:22 - Scenario 3: the online shop owner
17:53 - Scenario 4: the YouTube creator
18:22 - Where AI falls short (and why you stay in the loop)


And if you want the prompts used in the video, they're here:
https://benstevenson.notion.site/Loop-Engineering-Prompts-37bbe0f4c24f80f09b92f3884c109440?source=copy_link

I also put together a free guide with 10 loops you should get started with. It includes copy + paste prompts, instructions, and advice on which ones work best for your use case. You can get it for free here: https://www.sciencebased.ai/resources/loops