Make money doing the work you believe in
I fired ChatGPT.
Two years of my professional thinking was sitting on Sam Altman's servers. I just took some of it back.
Working drafts, decision frameworks, late-night ideas. I had given OpenAI a free training corpus on how I think. But after the “Department of War” deal and the political donations and the commitment to shoving ads into my AI relationship, I no longer trusted the people holding it. So I left.
The biggest lessons weren’t in the firing but in the exit interview.
I asked ChatGPT for a portrait of me before I closed the door, and Chat was SO excited to help me leave! What came back was a document I would never have written about myself. Ten things in total. Two I'll share right now.
1️⃣ I "often underweight operational friction or execution time." Translation: I underestimate how long things take. My team has been telling me this for years. I needed a chatbot to make me hear it.
2️⃣ I have a "low tolerance for performative enthusiasm." A tool built to be agreeable had learned not to bullshit me. That is more than some human collaborators have figured out.
The companies behind these tools have years of intimate data on us. Most of us have run the audit on THEM. Almost none of us have run it on OURSELVES.
Yes, applying an AI to read through your business data or medical info or refrigerator can be valuable. But remember: the most consistent thing they are reading is YOU.
There are many reasons to leave ChatGPT (and other LLMs). My team and I created an “AI Go Bag” to help you through it. In our case we migrated to Claude but that could change. What will sustain is the new practice of digital sovereignty and self-knowledge we now have practice at.
You can get the AI Go bag delivered to your inbox when you subscribe to this Substack.
Question for you right now: When did you last ask your AI what it knows about you?
