Prompting Skills playbook/Module 6 · Thinking with AI

Day 59 / 100 · Thinking with AI · Sycophancy and how to counter it

It wants to agree with you.

Models lean toward telling you what you want to hear. Knowing that is half the fix. Neutral questions are the other half.

Leading

Neutral

“Isn't this a great plan?”

“What are the three biggest risks in this plan?”

“I think X is right. Agree?”

“Give me the strongest case for and against X.”

“This is my best work yet.”

“Review this as a tough editor would.”

“You're wrong.”

“Show your evidence. Only change your answer if it's actually wrong.”

Source: arXiv ↗

Ask neutral questions. Get honest answers.