Daily principle
One short prompt-first habit to notice before work starts.
Founding beta · 15 min/day
Stop opening AI before your first thought.
Train a human-first AI workflow in 30 days: score your defaults, run one daily cognitive rep, log your AI use, and leave with rules you can keep.
Based on the published Anti-AI Brain protocolStarts with the 2-minute Anti-AI Brain ScoreNo payment collected on this beta page
Think before you prompt. Write before you paste. Decide before you ask.
I am not anti-AI. I am anti-dependence. I use the machine. I do not give it the first move.
Who it is for
Built for AI-heavy knowledge workers.
Founders, operators, writers, creators, engineers, consultants, researchers, and product people who already use ChatGPT, Claude, Gemini, or similar tools every day.
You open ChatGPT or Claude before you finish one rough thought of your own.
AI makes you faster, but you trust your own writing, recall, or judgment less.
Your voice gets flatter after too many model-assisted drafts.
You need AI for work and have no interest in quitting it.
You want rules, reps, and a score, not another vague digital-detox promise.
Daily rhythm
The beta is built around a 15-minute loop.
The first version should be easy enough to complete on a workday and concrete enough to show whether the protocol changes behavior.
Cognitive drill
A first-thought, first-draft, recall, or decision rep before AI enters.
AI-use log
Name what you outsourced, what you kept, and what needs a stricter rule.
30-day outcome
From reflexive outsourcing to deliberate AI use.
You do not quit AI. You move it later in the chain. The first thought, first draft, and first decision come from you again.
Attention
Put the pause back before the prompt
Track the impulse, delay the handoff, and rebuild the first human rep before the machine enters the loop.
Memory
Store substance instead of pointers
Use recall, handwritten notes, and end-of-day reconstruction so the work lives in you, not only in a tab.
Reasoning
Draft before you paste
Write the ugly first pass yourself, then use AI as critic, challenger, and falsification partner.
Judgment
Decide before you ask
Make the call in human language first. Let the model pressure-test it after responsibility is already yours.
How it works
Four weeks, one re-entry system.
The beta uses minimum viable reps for busy days and a full protocol for days when you can load the circuit harder.
Baseline
Take the Anti-AI Brain Score, name your weakest circuit, write your current AI defaults, and set the challenge rules.
Attention
Count AI impulses, run the five-second pause, and move one repeated prompt category back into human-first work.
Memory
Rebuild recall with daily three-paragraph reconstruction, handwritten carry-forward notes, and no-search summaries.
Reasoning and voice
Draft, argue, disagree, and use AI for Socratic challenge instead of first authorship.
Judgment and re-entry
Create your AI-use policy, relapse map, Day 30 manifesto, and Day 45 / Day 90 maintenance plan.
What you get
A container, not more content.
The value is the execution layer: pacing, daily check-ins, repeatable exercises, and a clear path from Day 0 score to Day 30 operating rules.
- Daily human-first reps in minimum and full modes
- Before / after Anti-AI Brain Score
- AI impulse log and daily completion check-in
- Weekly review prompts and failure-mode rescue
- Personal AI-use rules: what to outsource, audit, protect, and never delegate while learning
- Day 30 manifesto and 30 / 45 / 90-day maintenance plan
Book vs challenge
Understanding the protocol and completing it are different products.
The book can make the problem obvious in a weekend. The challenge exists to move the first rep back into your hands for 30 days in a row.
Explains the risks Trains the daily defaults
Gives the model Gives reps, check-ins, and pacing
Can be read passively Requires a human-first action every day
Shows what to protect Produces your personal AI-use rules
First aha
Start with the reps that change the sequence.
The beta does not start with theory. It starts with tiny friction before the machine receives the work.
Three bad bullets before prompt
Before asking AI for an answer, write three rough bullets of your own. Bad is allowed. Outsourced first thought is not.
Critic only
Paste your draft and ask the model to challenge, falsify, or pressure-test it. No rewrite request until your own draft exists.
Decision in one sentence
Write the call you would make without AI. Then let the model attack the tradeoffs, not choose the direction.
Join the founding beta
Get early access when the first test opens.
Leave your email if you want the founding beta. You will get one challenge confirmation email, the Score link, and the first invitation when the beta opens.
No payment is collected here. This page tests demand before the full paid challenge gets built.
FAQ
Quick answers.
Is this anti-AI?
No. The challenge is for people who keep using AI. The rule is sequence: think first, prompt second, use the machine as a sparring partner instead of a replacement nervous system.
Do I need to read the book first?
No. The book gives the full map and research base. The challenge gives pacing, accountability, and the daily reps. Readers will recognize the protocol, but the first beta is built to stand on its own.
How much time does it take?
The minimum track is 20 to 30 minutes a day. The full track can take up to 90 minutes. The point is not austerity. The point is restoring the human-first rep before AI use.
Is the beta paid?
The first page collects early-access demand. Founding beta options are expected to include a self-guided track and a guided cohort, but no payment is collected on this page yet.