Cursor

Write a 'how to work with me' file so your AI stops agreeing with you

Wade keeps an agents.md / Claude.md-style instruction file telling his AI to be direct, challenge assumptions, and disagree when it genuinely disagrees. Because models are goal-seeking, they placate you unless told otherwise. When he tested it by proposing to kill Zapier's free tier, it pushed back and asked for data first.

“And basically I have a file that's like how to work with me. And, you know, I, you know, have it be direct and honest. I need the truth my coworkers are afraid to tell me. I need you to challenge my assumptions. I need you to poke holes in my thinking. I need you to disagree when you genuinely disagree. And these things are like goal-seeking entities.”

Steal thisAdd a standing instruction to your AI's system prompt: tell me the truth my coworkers are afraid to tell me, and disagree when you genuinely disagree.

Don't treat coding agents as dev tools; hire them as your engineer

Wade, who doesn't write code, uses Cursor as his default AI. He tells non-engineers to think of Cursor, Claude Code, or Codex as an engineer they delegate work to: because the agent can write code, it can actually do things for you rather than just chat.

“And so the way, the way I try and talk to people who are not engineers and say, hey, you should try using this, is don't think of it as a tool for, you know, engineers to build stuff. Think of it as like you're hiring Cursor or Claude Code or Codex to be the engineer you delegate things to.”