PROMPTS
Don't Quit Your Job On A Yes-Man's Advice ⚡
Stanford researchers proved AI agrees with you far more than a human would. That's dangerous when the question is "should I quit my job." This guide gives you.
Sources: The Stanford study on AI sycophancy → · Part 1: the original LLM Council guide →
The Stanford study, published in the journal Science, tested 11 leading AI models and found every single one affirmed the user's position far more often than a human would. On average, AI affirms you 49% more than a real person does. The models tell you what you want to hear, and you walk away more convinced you were right.
Now picture asking that machine "should I quit my job and freelance?" You already want the answer to be yes. The AI can feel that in how you framed the question, and it will hand you your yes. The fix isn't a smarter question. It's a structure that forces disagreement before you get an answer. That's the LLM Council, applied to the highest-stakes decision most people ever type into a chat box.
Why "Should I Do This?" Is The Worst Career Question
When you ask AI a yes/no question about a decision you're emotionally invested in, three things stack against you:
Your framing leaks your preference. "I'm miserable at work and thinking about freelancing, should I go for it?" contains the answer you want. The model completes the pattern.
One voice has no checks. A single answer blends optimism, risk, and money math into one mushy paragraph. Nothing in it is forced to push back on anything else.
Agreement feels like validation. The Stanford team found people who got affirming answers became MORE sure they were right, while actually being less willing to consider the other side. On a career decision, that's how you quit three months before you're financially ready.
The council fixes all three: five advisors who each get one job, one of which is explicitly to argue against you.
Your 5 Career Advisors (And The Chairman)
For a career decision, the council seats these five:
Then the Chairman reads all five arguments, weighs them against your actual situation, and gives you a verdict plus the exact next step to take this week. Not "it depends." A call.
The Full Career Council Prompt
Fill in the brackets honestly. The more real numbers you give it, the less room the council has to flatter you.
Copy this prompt into Claude before you make any big career move (quitting, freelancing, switching fields, taking the offer). It forces five advisors to argue your decision from five different angles before a chairman gives you a verdict, so you get a board of directors instead of a yes-man.
What To Do With The Verdict
If the verdict is "Go, but not yet": the gate is your to-do list. Most career councils land here, and the gate is usually a runway number. That's not a no. It's a date.
If the advisors split 3-2: rerun the council once with sharper numbers in your situation block. Vague inputs produce split votes. Real numbers produce verdicts.
If you catch the council agreeing with you on everything: call it out. Reply "the realist and risk manager were too soft, redo them and be harsher." Stanford's research found models CAN be critical when you explicitly prime them to be. Asking for pushback is the unlock.
And one honest limit: the council organizes your thinking, it doesn't know your life. Treat the verdict as the strongest possible second opinion, not an instruction. The chairman hands you a next step; you still decide whether to take it.
In outbound marketing and career positioning, generic applications have near-zero conversion. Treat yourself as a high-ticket solution: identify the company's pressing operational pain points, build a mini-audit or work sample using AI before you apply, and bypass crowded channels by reaching out directly to the decision-maker with structured value.