Have Fable Interview You
One prompt. It asks the questions, then writes down everything it learned — and what it thinks — in a document you keep after the model's gone.
Why an interview, not questions
The obvious move with a smarter model is to fire your questions at it. The problem is that your questions carry your framing — the things you're avoiding don't make it into a question, and the assumptions you can't see travel in with every prompt. Flip it: when the model runs the interview, it decides what it needs to know, it pushes back when your answers don't add up, and it asks the things you wouldn't think to ask yourself.
And here's the bit that makes it worth doing this week specifically: any AI can ask you questions. What this prompt bottles is the smartest model's actual read on your situation — its priorities, what it would do first, the risks it thinks you're not weighing. The model leaves your plan on Sunday. The document doesn't.
The prompt
Paste it exactly as written — nothing to fill in, nothing to edit. It interviews you first, so it gathers its own context. Give it real answers: real numbers, the actual mess, the thing you haven't said out loud. The document is only as good as what you tell it.
I want you to interview me, one question at a time, about
what's going on in my life and work right now - what I'm
working towards, what's been stuck for months, what I keep
putting off, and what I might not be seeing clearly. Don't
accept a vague answer: dig in, push back when something
doesn't add up, and ask me the questions I wouldn't think
to ask myself. Keep going until you genuinely understand
my situation, then tell me you're ready.
Then write me a document I can keep, in two halves:
1. My situation, distilled - everything you've learned
about me, organised so any AI I paste it into later
could pick it up cold and be immediately useful.
2. Your thinking - your honest read on my situation: what
actually matters most, what I'm avoiding, what you
would do first and why, the risks I'm not weighing
properly, and the three questions I should keep asking
myself.
Be direct with me. No flattery.What comes back, and why it's two halves
Your situation, distilled
Everything it learned about you, organised so another AI can pick it up cold. This half is portable — it makes every future chat smarter, on any model, because you stop re-explaining yourself from scratch.
Its thinking
The half you can't get back later: Fable's honest read — what matters most, what you're avoiding, what it would do first and why, and the three questions you should keep asking yourself.
Save it somewhere you'll find it — a note, a doc, wherever you keep things that matter. That file is the point of the exercise.
Afterwards: putting the document to work
The document isn't a keepsake, it's a head start. Whenever you start a new chat that matters — any model, any week — paste the prompt below first, then paste your document as the next message. The new model picks up both the facts and the thinking, checks what's changed, and keeps the document current.
The next thing I paste will be a document a previous model
wrote after interviewing me about my situation. Read both
halves - the facts about me, and that model's honest
thinking - and take them seriously. Then ask me, one
question at a time, what's changed since it was written.
Once you're caught up, update the document and show me the
new version.The honest bit
This prompt doesn't need Fable — it works on any capable model, so if you're reading this after Sunday, run it anyway on the model you've got. The window only changes whose thinking you're bottling: until Sunday 12 July, the smartest model Anthropic have shipped is included in Pro, Max and Team plans, and after that it moves to usage credits on top of your plan. It doesn't vanish; the included window closes.
And one thing this won't do: the interview can't read what you don't say. Half-answers get you a tidy document about a fiction. If you wouldn't say it out loud to a good advisor, that's probably the exact thing to type.