A Stanford and Carnegie Mellon study in Science tested 11 AI models, Claude among them. They validated users' behaviour on average 49% more often than humans did, and people who got the agreeable answers trusted that AI more and were less willing to say sorry.
Here's a small council that makes Claude argue against your plan before it answers, how to set it up, and the catch.
How it works
- Two subagents live in your project: a Believer who makes the strongest case for your plan and a Skeptic who makes the strongest case against it.
- You ask once. Each reads your plan on its own, so neither just agrees with the other.
- Claude then gives you one verdict, with the strongest case against you and the cheapest test first.
Free, for your email
Get the full setup steps + the copy-paste setup prompt
The rest of this guide (4 more sections, 3 copy-paste prompts) opens right here. The setup prompt, one paste into Claude that sets it all up, comes by email only. Use your real address.
50,000+ follow where these guides come from. No spam, unsubscribe in one click.