ADHDThe Postcard Rule: Yap Freely Without Oversharing
Before you yap into an AI, ask one question: would you write this on a postcard? A postcard travels through a dozen strangers' hands before it reaches a mailbox. Anyone who touches it can read every word. That's roughly how private a chat with an AI actually is. (Settings vary — some tools, like Shimmer, publicly promise not to train on your data — but the postcard rule keeps you safe everywhere.) Nobody's necessarily snooping, but you don't fully control where those words get stored, how long they sit there, or what happens to them later. Treat the chat window like the back of a postcard, not like a locked diary.
The postcard rule gets you talking more, not less. Once you decide what's off the postcard, everything else on it is fair game. You can yap about the ugliest, messiest, most frustrated version of your day, because you already set the fence before you opened your mouth. The fence is what buys you the freedom.
The names go first
The easiest way to keep a yap postcard-safe is to swap real names for placeholders before you start talking. Your business partner becomes Person A. The client who's been ghosting you becomes Company B. The project that's three weeks late becomes Project X. You lose nothing by doing this. The AI doesn't need to know it's actually your cousin, or the deal in Denver, to help you think through what to do next. It needs to know a person is late, you're frustrated, and you have a call with them Thursday.
Pick your placeholders once and reuse them across a session, or even across weeks, so the AI can follow the thread without you ever typing anything that identifies a real person, a real company, or a real project. Person A, Company B, Project X. That's the whole system, and it takes about ten seconds to set up before you start yapping.
What never goes on the postcard
Four categories stay off, no exceptions: medical details (diagnoses, medication names, dosages), legal details (case numbers, attorney names, court dates), money details (account numbers, exact balances, Social Security numbers), and anything you wouldn't want a stranger reading over your shoulder at a coffee shop. You can still yap about how a diagnosis made you feel, how the legal stuff is keeping you up at night, how the money situation is scaring you. Talk about the feeling and the shape of the problem, not the identifying specifics. "I'm scared about a health thing" is postcard-safe. The name of the condition and the dose of the medication is not.
Same goes for exact numbers tied to you personally. "I'm short a few thousand dollars this month" tells the AI everything it needs to help you think through options. Your actual account balance tells it nothing more useful, and it's the kind of detail you don't want sitting in a chat log anywhere.
The postcard rule
What's safe to yap vs. what stays off the postcard
You're not hiding your feelings. You're just changing the names of the people you feel them about.
The fence is what lets you yap more
Once the rule is automatic, you share more, not less. When you're not stopping mid-sentence to wonder whether you should say his real name, you talk faster and longer. You stop editing yourself in real time, which defeats the whole point of yapping in the first place. Set the fence once, before you start, and nothing inside it needs a second thought while you're mid-rant.
This matters even more for an ADHD brain. Stopping to self-censor mid-thought is exactly the kind of extra step that derails a train of thought that was already hard to catch. A 2026 co-design study of 20 ADHD university students and 5 experts, published at CHI 2026, named emotional regulation as one of the core things ADHD adults need support with when working alongside AI. Our read: You can't regulate an emotion you're too busy editing. Decide your fake names and your four off-limits categories before you open the mic. Then just talk.
AI can be wrong, so keep the leash short
A December 2025 review in BMC Psychology put AI-generated ADHD executive-function plans in front of 27 clinical experts. They found the plans were theoretically well-grounded but often too generic to survive contact with a real week. Their conclusion was that AI should complement human judgment, not replace it.
That study is about plan quality, not privacy, but the lesson carries over. A tool that gets the plan wrong sometimes isn't a tool you hand your Social Security number to, either. Keep the sensitive stuff with your actual doctor, your actual lawyer, your actual bank. Keep the postcard-safe version, the feelings, the shapes, the fake names, with the AI. AI can support the thinking. It doesn't get to hold the paperwork.
Try it on your next yap
Before your next session, pick three placeholders you'll actually reuse: one for the person you complain about most, one for the business or client, one for the project. Then yap like normal. Talk about how stressed you are, what's stuck, what you're avoiding, what you wish someone would just handle for you. Leave the real names, the diagnosis, the case number, and the account balance off the postcard. Everything else is fair game, and now you know exactly where the line is, so you can stop thinking about it and start talking.
Take the whole system with you
The free 43-page ebook: the brain science in plain words, the four steps, the master prompt, and your first seven days.
Prefer it live? Doug teaches this in libraries and meetups — book a 90-minute class →