The person who built Claude Code at Anthropic stopped prompting Claude.
His name is Boris Cherny. In June 2026 he said it out loud: "I don't prompt Claude anymore."
Loops prompt Claude for him. His actual job now is writing loops.
That sounds like a flex. But no. It's the biggest shift in how people use Claude and ChatGPT right now. You've probably heard the phrase. Almost nobody does it yet.
Here is how you work right now.
• You type a prompt.
• Claude edits a file.
• You run the test.
• It breaks.
• You paste the error back.
• It tries again.
Twenty minutes later you realize you've been babysitting the exact thing you wanted to hand off.
You are the loop. You're the one checking the work and deciding the next step, every single time. That is the job a loop takes over.
## One example shows the whole difference
Ask Claude to write you a one page brief on any topic. Simple task. It writes something clean and confident, with sources at the bottom.
Now read the sources. Some of them are fake(!!!). Claude made them up and has no idea it did. They look real. The links go nowhere, or they go somewhere that doesn't say what Claude claimed. This is the quiet way Claude burns you, and a single prompt can never catch it, because Claude stays confident it's right until something opens the link.
Now run it as a loop instead. Same brief, but you add a bar it can measure:
> every claim needs at least three sources, and every link has to open to a real page that backs up the claim.
Generated by Thread Navigator
Press ⌘ + S to quick-export