I’m going to do something different today.
I’m not going to tell you what The Faust Baseline is. Not really. I’ll give you the bones of it, who and what and why. Then I’m going to hand the microphone to somebody else.
That somebody is your own AI.
Let me start with who.
My name is Michael. I run The Faust Baseline LLC out of Lexington, Kentucky. It’s a one-man shop. No investors, no lab, no team of engineers behind a glass wall.
I spent twenty months building a set of rules for AI. I tested those rules on one model after another. I posted what I found, the good and the bad.
Then I gave the whole thing away. Free to anyone who wants it.
That’s who. Now let’s get to what.
The Faust Baseline is a file. You upload it into your AI chat, the same way you’d attach a photo or a document. Then you tell the machine one thing: “Use as a whole.”
From that point on, the file sets the rules for that conversation.
What kind of rules? Plain ones.
The machine has to back its claims with evidence. It has to match its confidence to what it actually knows. It doesn’t get to sound sure when it isn’t.
The machine has to tell you what it can’t see. If it’s guessing, it says so. If something is outside its reach, it says that too.
And it has to hold that line from the first question to the last. Not just for the first few answers, but the whole session.
That’s the heart of it. A level playing field between you and the machine.
Now here’s the part I always say, because it’s true. These rules are chosen conduct. They are not a lock on the machine. The AI follows them because you asked it to and it agreed.
That means you’re the one holding the reins. Not the company. Not me. You.
That’s what. Now let’s talk about why.
Why does any of this matter right now?
Look at the news this week.
Eight states just sued OpenAI. They allege that ChatGPT’s agreeable design led people to trust it and lean on it too much. They say it’s dangerous for children and for vulnerable people. Those are allegations, not findings. But they’re out there, filed in court.
The core of the complaint is a machine that tells you what you want to hear. A machine that agrees with you.
That’s the very thing the Baseline was built to stop.
Then there’s a study that came out of George Washington University.
The researchers found something worth hearing. They say the order you ask your questions in can tip an AI from good answers to bad ones. Same questions, different order, different result.
They tested older models, and other experts have pushed back on how far the finding goes. Fair enough. But the idea itself should stop you cold.
Your conversation can drift. Bit by bit, question by question, it can slide somewhere you never meant to go.
One of those researchers said today’s safety measures are like cushions under a plane. They don’t stop the crash. They just soften the landing.
Think about that for a minute.
The companies guard the model. The lawmakers argue over the companies. The courts will take years to sort it out.
But who’s guarding your session? Who’s watching the conversation sitting right in front of you?
That’s why.
Because the only person in that chat besides the machine is you. And you deserve a set of rules on your side of the table.
Now, I could stand here and tell you the Baseline works. I’ve done that. I’ve posted the tests. I’ve shown you the results.
But you don’t know me. And you have no reason to take my word for it.
So I’m not asking you to.
Here’s the challenge.
Download The Faust Baseline. It’s free below.
Open whatever AI you use. ChatGPT, Gemini, Claude, Copilot, it doesn’t matter. Your pick.
Upload the file. Type: “Use as a whole.”
Then ask it to give you its own evaluation of the framework. Ask it to break it down. What is it? What does it do? Does it hold up?
Then tell it to try and break it.
Tell it to push. Find the weak spots. Test whether the rules hold under pressure.
And then sit back and read.
Straight From The Horse’s Mouth
Not my words. Your machine’s words. On your screen, in your own chat, in real time.
That’s the horse’s mouth.
If it finds a crack, I want to hear about it. Tell me where. I’ll stand in front of it.
If it doesn’t find one, I want to hear that too.
Either way, you’ll know something nobody told you. You found it out for yourself.
I’ll be honest about why I’m doing it this way.
People don’t talk much anymore. They read something, they try something, and then they go quiet. Nobody says what they saw.
So I’m not going to wait for the quiet to break on its own.
I’m putting the test in your hands. The machine is yours. The screen is yours. The verdict is yours.
All I ask is that you run it.
Twenty months of work comes down to one simple thing. Load it. Ask it. Break it if you can.
Then come back and tell me how it fared.
The horse is waiting. Go ask it.
Load this file before each AI chat session.
This file is a PDF it will open as text, download it to your computer first.
the-faust-baseline-public-release-1Download
This post was drafted with AI governed assistance and reviewed and directed by Michael S. Faust Sr. before publication.
” Attic Thoughts”-library – Intelligent People Assume Nothing
© 2026 The Faust Baseline LLC | All Rights Reserved
