Michael Stelzner said something on his show that made my ears perk up. He was talking about those one-on-one calls where he feels like he's in "a little zone" and "gold comes out." He wishes a camera had been running, because he already knows it would be content.

Think about a call like that. Someone brings you a problem, you start talking it through, and somewhere in the conversation you find an answer you couldn't have prepared in advance. Then the call ends, and all that thinking is scattered through an hour of conversation.

He was describing the blind spot I'd come on to talk about.

When you describe yourself in an interview or a prompt, you leave out decisions you make without thinking. The version of you solving a problem with a client is the one AI rarely gets its hands on.

Our AI Explored conversation is above, and I've expanded on the method below.

A prompt that literally changed my life

I spent the first 16 years of my career in corporate marketing, working in pet healthcare and on baby brands. During COVID, I started consulting for smaller businesses and found that I loved the one-on-one work.

After COVID, I spent $10,000 getting certified in a messaging framework and came home ready to use it. Around the time I started using ChatGPT, someone sent me a prompt that recreated what I'd just paid to learn.

The painful feeling in my stomach was the 10k burning its way out.

Once I was mentally recovered I took a long, hard look at my career. If AI (which was in its VERY early stages) could already do this, what does that mean for me as a marketer in the next 3, 5, 10 years?

I knew something had to change and I better figure this puppy out. I started spending 10 or 12 hours a day learning, playing and experimenting.

Before long, I realized that 80% of my marketing consultations were now spent on me helping people understand AI and so I ripped the band-aid off and went all in.

I started doing one-on-one AI consulting and implementation calls. They were going well, but when I thought about "why" they were going so good I couldn't really put my finger on it.

In an attempt to practice what I preached, I put AI on the job and had it dig into all of my transcripts with clients.

The first, amateur prompting pass just uncovered that people kept repeating "I just thought about AI differently".

Interesting...I went deeper.

I started looking at the decisions that I made, the repeated patterns of how I introduced a problem then bridged to a solution, how I was able to spot (in advance) when people were getting overwhelmed and would scale things back and simplify.

I. Was. Hooked.

And if I was getting so much out of it, wouldn't other people too?

I started looking for those patterns in other people's transcripts and reading them back to them. They recognized something they did but hadn't found words for.

And that's how, in a very abbreviated version, Cognitive Fingerprint started.

🧠 Join CF Lab

An AI interview captures what you can explain

An AI interview can help you explain your work, but some of what makes you good has become so automatic that you don’t even realize you’re doing it. That’s unconscious competence. I was spotting overwhelm and simplifying the conversation before I’d recognized either as part of my process.

With a transcript, you can slow that moment down. The client says something, you change your approach, and the conversation goes somewhere else. You can read back through it and ask what you picked up on.

That was what pulled me deeper into my own calls. I could start to see how I was doing the work people valued. Then I could open another transcript and look for it again. You can ask AI to help with that search and show you the passages behind each pattern it suggests.

You can spot a bad headline before you can explain why

Michael described this as a copywriter. He can look at a headline, feel that something's off, and rewrite it. Explaining why he changed it is harder.

Philosopher Michael Polanyi wrote about knowledge we have trouble putting into words in The Tacit Dimension. His phrase was "we can know more than we can tell."

[Image placeholder: A transcript can reveal decisions you might leave out when describing your process.. We’ll generate and insert this.]

Imagine having a handful of Michael's original headlines next to his rewrites, with his comments alongside them. We'd have somewhere to start looking. Does he keep replacing broad promises with something the reader can picture? Does he pull a detail out of the article and put it in the headline?

Those are questions we'd test against the edits. As we found recurring choices, we could start describing what his eye catches when a headline feels off. A judgment that was hard to explain would have examples we could point to.

I call those recurring decision patterns decision DNA. They're useful context for AI because they show what you pay attention to and how it affects what you do next.

Collect calls where you're solving a problem

Start with conversations where you're responding to someone as you go. Client consultations and coaching calls are useful because the other person's questions can take you somewhere you hadn't planned. Sales calls, brainstorms, and employee one-on-ones can give you other examples.

Picture a client call where you arrive ready to explain something and realize halfway through that you need to back up. Maybe you ask them to show you what they've tried. Maybe an example gets the conversation going again. Those little adjustments are worth looking at.

You need the transcript, including what each person said. A meeting summary might reduce that whole exchange to "discussed implementation." I'd want to read what happened just before you changed direction.

I use Granola, which transcribes meetings without adding a bot to the call. If you already have transcripts from another meeting tool, start with those. Get permission to record and use the conversation before collecting someone else's words.

You can also capture your own thinking out loud. I use Wispr Flow when I start projects in Claude Code or Codex, and when something in a podcast makes me want to stop and work through an idea. Those notes give me material I might lose if I waited until I was ready to write about it.

Start with where your conversations happen, then check how to get the full transcript out. You might already have what you need in a tool you're paying for.

I suggest starting with three to five transcripts from different situations, because I want to see a decision at least three times before I treat it as a recurring pattern.

I've processed more than a thousand files for one person. You can start much smaller. As the same patterns keep appearing, new transcripts become useful for checking whether anything has changed.

Save each transcript as a readable text file.

At the top, identify who was speaking and what kind of conversation it was. A client coaching session and an internal brainstorm put different demands on you. That context helps the model interpret why you responded a particular way.

Keep the transcript files in one folder so you can find them again when you compare patterns across calls.

Four layers I pull from a transcript

The names for these four layers sound a little academic. It helps to keep a real conversation in mind as you read them. Think about one where you knew what you wanted to explain, then had to adjust as you went.

1. Declarative knowledge

What you know and can describe. Your account of what you do belongs here, including the expertise you'd put in a professional bio.

2. Procedural knowledge

How you do it. A transcript may show the sequence of questions you ask or the steps you take to work through a problem.

3. Conditional knowledge

When and why you choose a particular approach. What did the client say that made you ask another question? What would have led you to recommend something else? This is where I look for decision DNA.

4. Metacognitive knowledge

How you examine your own thinking. You might question an assumption or reconsider how you're approaching a problem. I've asked people to name the mental models they use, then compared their answers with what appeared in their transcripts. The two accounts can differ.

In my calls, the interesting question became what made me simplify at that particular moment. I could teach someone the steps of an AI workflow. Understanding when to slow down meant paying attention to the person trying to learn it.

That's why I spend time with the conditional and metacognitive layers. I want to understand the choices tucked inside an ordinary conversation, including the ones I made too quickly to notice.

Use the Four Layers prompt linked below with one transcript, then repeat with the next few calls. Compare the suggested patterns across them and keep the examples that support each one.

Turn those patterns into a file AI can use

After working through an archive of calls, a full analysis can run 20 to 30 pages. I trim that into a fingerprint file you can give AI as context when you start a project. It describes your recurring decisions and includes examples, so the model has something specific to refer to when it helps you.

For a first version, make a text document with the patterns you've checked against your calls. Give each pattern a short name, explain the decision it describes, and include the transcript passages that helped you recognize it.

Paste that document into a new AI conversation with your next request and ask it to use those examples when helping you.

Read its response to see whether it understood the pattern and applied it appropriately.

You can keep that file and reuse it when you change AI tools. Review it as your work changes, too. An approach that suited one kind of client may need adjusting for another.

There's a personal reason I love this work, too.

When I started reading patterns back to people, they recognized things they'd been doing without having words for them.

I knew that feeling from looking through my own calls, and now I was getting to share the experience.

Imagine trying to explain your work to a potential client with a few of those examples in front of you. You could show how you spotted a problem and decided what to do about it. Those examples can help you build an offer around how you work or put your approach into a facilitation guide.

With teams, comparing people's patterns can help you decide who should lead a project or how to explain an idea to a colleague. If a partner tends to ask for data while you tend to explain through stories, you can account for that when you prepare for a conversation.

A transcript gives you evidence to check. If a description of your thinking doesn't fit, go back to the exchange it came from. You should be able to recognize what the model is describing and question it when you don't.

Update & Prompt!

One announcement and one freebie below…

As a thank you to all my paid subscribers, I am adding a TON of CF resources to a Skool community. You paid membership to Substack ALSO includes the community (who doesn’t love a good BOGO?!).

Cognitive Fingerprint™ Lab is a Skool community built on the methodology you just read about.

Inside There Are 5 Pillars

1. The Fingerprint: Why you can’t see your own expertise. The science that explains why the thing you do better than anyone else feels like nothing special to you.

   This is where most people realize they’ve been sitting on a goldmine and treating it like gravel.

2. The Map: The full territory sitting inside one recorded conversation, charted. Most people run ONE extraction, get ONE result, and stop. They think they got everything. They didn’t. This pillar shows you what you’re actually looking at when you record a session…

   And how to pull out layers most people never even know exist.

3. Extraction: The prompts, workflows, and tools you run against your own calls. The Exception Finder below? That’s the shallow end. Inside the Lab, you get the full toolkit. The ones that find corrections you didn’t know you made. That spot the belief systems hiding in your repeated language. And turn hesitation into a teachable decision framework.

   This is where your invisible judgment becomes something you can actually USE.

4. Your Method: Turning repeated patterns into a named method you can sell. Real intellectual property you are already sitting on. Or a SOP your team can run without needing you in the room. Because once you’ve captured your judgment, you can SCALE it.

   You can train people using the exact distinctions that make YOU different.

   You can hand off work you used to think only you could do.

   And you can finally stop being the bottleneck in your own business.

5. The Guide Position: Pricing, positioning, and content built on what you found.

   Once you know what you actually do that nobody else does…

   You can TALK about it.

   You can write about it in ways that make people say:
   “Holy moly, this person gets it.”

   You can price yourself based on the value of that judgment…not the hours it takes to apply it. And you can position yourself as the person who solves a problem most people don’t even have language for yet. This pillar is about turning captured expertise into market position. The kind clients pay premium rates for.

Membership includes both this newsletter AND the Lab. Whichever door you come in through, you get both.

Since these two platforms don’t really talk to each other the quickest way in is vis the Skool link below.

🧠 Join CF Lab

OK OK…now for your prompt…

4 Layers Prompt

Pick a call you remember enjoying, one where you helped someone work through something and came away pleased with how it went. Get the free Four Layers prompt, enter your email to have it sent to you, and try it with that transcript.

When a suggested pattern catches your attention, read the conversation around it. See whether you remember why you asked that question or changed your approach. You might find yourself opening another call to see if you did it there, too.

A waste-management owner spent 30 years learning to “read” dumpsters.

Sounds weird, right?

But he could drive past one and instantly know if there was a deal there. The trash inside told him something. The way it was painted told him something else. The rust patterns. The placement. All of it. His new salespeople? They just saw a dumpster.

And when it came time to train his staff, he couldn’t explain what he was seeing. It was so natural at that point, all so effortless. It was invisible to him, so he never thought to explain it in the training manuals. “Reading” a dumpster felt like doing his job, nothing else.

Thirty years of market knowledge like this can disappear because nobody ever called it knowledge in the first place.

The first two pieces in this series explained why expertise goes invisible and why AI makes that a real problem now. This one is about what to do before it disappears for good.

🧠 Join CF Lab

Why Asking For Someone’s Process Never Works

Ask an expert how they do their work and they’ll usually hand you something that sounds pretty nice. It will give all of the “official” steps, maybe a framework, and the rules they remember learning. What most SOPs look like.

That stuff has value. But it almost never contains the moments where their skill actually changed the outcome.

The real material shows up when the plan stops working.

When the expert hits a wall and breaks from the normal process.
When they catch something that wasn’t in the brief.
When they fix someone’s work without being able to say exactly what rule they just applied.

That’s why transcripts, corrections, hard cases, lost deals, rewrites, and odd calls hold so much more than a normal interview. The expert has stopped explaining. They’re applying the method live, in real time, fully inside their zone of genius.

The Evidence Judgment Leaves Behind

Judgment leaves evidence behind. You just need to know where to look.

1. Corrections

When an expert changes an AI’s output, a teammate’s work, or their own first draft, put the two versions side-by-side.

The gap between them is the standard.

My client could tell when AI output read like a telegram instead of a magazine. That distinction lived in her head until she made the correction.

One fix like that is just an anecdote.

Line up enough corrections and sort them by what changed and why, and a pattern starts to show up that someone else could actually learn from.

Every time you change something, you’re applying a standard. The standard is the valuable part.

The edit is just the evidence of it.

2. Rejected options

The option an expert turns down often says more than the one they pick.

Most people only write down what they decided and what they shipped.

Write down what got rejected too. And when. And why.

A list of final answers shows you a preference.

A record of good options getting cut shows you judgment.

That’s the real difference between knowing what someone chose and knowing how they actually think.

3. Exceptions

Every process eventually hits a case it wasn’t built for. The exception shows you what the expert thinks the process is actually protecting.

They break from it because they can see that following it here would give the wrong result, even though they still believe in the goal the process was built for.

The philosopher Donald Schön called this reflection in action. The situation stops matching the theory, and the practitioner adjusts on the spot.

That adjustment usually carries more real skill than the plan it just broke.

Gary Klein studied how firefighters, military commanders, and ER staff decide things under pressure. They almost never weigh options side by side. They recognize a pattern and move.

It’s fast because the experience runs deep, and the exception is where you finally get to see that depth.

4. Comparisons

Experts are always comparing the case in front of them to cases they’ve already seen.

“This looks like that one, except for this one thing.”

That one thing might be the whole decision.

The comparison shows you two things at once.

1. What the expert thinks is similar.

2. What they think is different.

The similarity tells you which bucket they put the case in.

The difference tells you what actually separates cases inside that bucket.

Most frameworks flatten this out. They name the bucket and lose the variation.

The variation is where the judgment actually lives.

5. Repeated language

My client had a phrase she used with certain patients.

She called it the “Museum of Faith.”

When I showed her that phrase back to her, she corrected it.

“Museum of Toxic Faith.”

At first I thought it was just a word swap, but I quickly realized it was a belief.

A belief about what faith and hope does when it keeps someone holding onto a future that’s never going to show up, the line between hoping and actually changing and about the moment encouragement turns into a way of dodging something the patient already knows.

She’d never said that belief out loud before.

It only showed up because she rejected a phrase that almost fit.

The words an expert keeps repeating, the images they keep reaching for, the phrasing they correct.

That’s judgment made visible.

You don’t need to interview anyone for this.

You need to notice what they say more than once, and what they refuse to let stand.

6. Moments of hesitation

A pause can show you two standards fighting each other.

The expert is weighing two readings that both have evidence behind them. Both could be right.

The hesitation is the second right before one of them wins.

That moment holds more useful information than the confident answer that comes after it.

The confident answer already settled the fight.

The hesitation shows you the fight itself.

If you only write down conclusions, you lose the reasoning.

The reasoning is the part someone else would actually need to learn.

How Do You Know You Actually Captured Something Real?

How do you know what you’ve captured is real, and not just a story that happens to sound good?

A captured pattern earns more trust under a few conditions.

• It shows up three or more times: One correction could just be a preference. The same correction across five different cases is probably a real standard.

• You can point back to the original words or decision: The pattern isn’t a summary someone made up. You can trace it back to something the expert actually said or did.

• The expert can correct your read on it: Show them what you found. They tell you whether it’s real. They might sharpen it. They might reject it outright. Either way, it gets better through their response.

• The pattern predicts what they’ll do next time: This is the real test. If you’ve actually captured something, it should help you guess what they’ll notice, reject, or care about in a brand new case. If it does that, it’s earned its keep.

• It keeps the contradictions instead of smoothing them over: Real experts hold two competing ideas at once. They know when a rule applies and when it should bend for another one. If your write-up makes someone sound simpler than they are, you probably lost something important.

This is what separates the real thing from a flattering personality summary or an AI-generated framework mill.

A pretty description description isn’t enough.

The interpretation has to survive getting tested against another real decision.

What a captured pattern can turn into

Three people I’ve worked with saw a different future once their judgment finally got written down.

• A founder pictured his team walking into client calls “super confident, not really worried if I’m on the call or not.”

  The patterns he’d never said out loud became something his team could actually learn. His private judgment turned into something the whole team could run on.

• Another pictured content built from the thinking inside his firm’s client conversations instead of generic SEO filler.

  Thinking that used to stay trapped in conversations turned into something he could publish. His firm could finally say what it actually knew.

• A third could start giving AI more than a vague adjective about tone.

  Her corrections could turn into real examples of her standards, examples a model could actually learn from instead of guessing at what “magazine” meant.

These are real results from early sessions, not a promise about what everyone will get.

This work is still new (well, new compared to legends like Gary Klein and Polany).

What I’ve seen so far is what people recognized and pictured once they saw their own judgment reflected back at them.

It doesn’t yet prove every one of those ideas got adopted for the long haul.

The evidence for what this can become keeps growing.

It isn’t finished yet.

What Capture Can’t Do

Some expertise only comes through doing something together, or watching it happen, or seeing enough cases in person.

Words won’t carry all of it.

Some judgment should stay private too. It might hold confidential material or a real edge that shouldn’t get handed out.

The goal here is a useful translation, not a perfect copy.

This loops back to the sapphire labs from Article Two.

The visiting researchers never turned into copies of the original technicians. They just got close enough to reproduce the result.

That’s the real bar.

You don’t need to know everything, you just need to get close enough.

The Beginning

I’m giving you a prompt below that’s one of the tools I’ve built over the past two years.

It’s part of a methodology I call Cognitive Fingerprint™.

And if you just ran it on one of your calls…

You probably found something you didn’t know was there.

Maybe a pattern you repeat without thinking, a standard you apply but never named or a break in routine that reveals what you’re ACTUALLY protecting when you break your own rules.

That’s just the start.

Cognitive Fingerprint™ Lab

These three articles are building toward something bigger.

And a change in what this newsletter is becoming. Because here’s the truth…

A newsletter can show you the territory.

It can give you a prompt like the one above and can prove the method works.

But it can’t give you the FULL toolkit or walk you through extracting patterns from YOUR calls.

It can’t show you how to turn those patterns into something teachable, scalable, or valuable.

And it can’t give you a place to actually DO this work… with others who are doing the same thing.

For that, you need the Lab.

Cognitive Fingerprint™ Lab is a Skool community built on the methodology you just read about.

Inside There Are 5 Pillars

1. The Fingerprint
   Why you can’t see your own expertise.

   The science that explains why the thing you do better than anyone else feels like nothing special to you.

   This is where most people realize they’ve been sitting on a goldmine and treating it like gravel.

2. The Map
   The full territory sitting inside one recorded conversation, charted.

   Most people run ONE extraction, get ONE result, and stop.

   They think they got everything. They didn’t.

   This pillar shows you what you’re actually looking at when you record a session…

   And how to pull out layers most people never even know exist.

3. Extraction

   The prompts, workflows, and tools you run against your own calls.

   The Exception Finder below?

   That’s the shallow end.

   Inside the Lab, you get the full toolkit.

   The ones that find corrections you didn’t know you made.

   That spot the belief systems hiding in your repeated language.

   And turn hesitation into a teachable decision framework.

   This is where your invisible judgment becomes something you can actually USE.

4. Your Method

   Turning repeated patterns into a named method you can sell. Real intellectual property you are already sitting on.

   Or a SOP your team can run without needing you in the room.

   Because once you’ve captured your judgment, you can SCALE it.

   You can train people using the exact distinctions that make YOU different.

   You can hand off work you used to think only you could do.

   And you can finally stop being the bottleneck in your own business.

5. The Guide Position

   Pricing, positioning, and content built on what you found.

   Once you know what you actually do that nobody else does…

   You can TALK about it.

   You can write about it in ways that make people say

“Holy moly, this person gets it.”

1. You can price yourself based on the value of that judgment…not the hours it takes to apply it.

   And you can position yourself as the person who solves a problem most people don’t even have language for yet.

   This pillar is about turning captured expertise into market position.

   The kind clients pay premium rates for.

Membership includes both this newsletter AND the Lab. Whichever door you come in through, you get both.

Right now, it’s $20 a month. That price is absurd. I know it.

But I know how amazing this stuff is and I’ve been only giving it out to my one-on-one clients.

I’ve seen founders extract their own decision frameworks and realize they can finally train their teams properly.

I’ve seen consultants pull positioning language straight from their own client calls and turn it into content that actually differentiates them.

I’ve seen people who thought they “just did their job” discover they had a methodology worth naming.

But I do want to reward the early-birds, so the price is going up.

At midnight tonight, September 5th, new memberships move to $25 (and will probably keep rising).

Anyone who joins before then keeps $20 for as long as their subscription stays active.

I was supposed to close the $20 price last night but I got delayed getting this article out.

So I’m leaving the door open until tonight at midnight, 9/5.

After that?

It’s $25 for everyone new.

And it’ll go higher from there.

Now, go find something you knew, but didn’t know, about how awesome you are.

The Exception Finder

The exception is the fastest place to catch your own judgment firing.

Every time you skip a step, change the order, or break a rule you normally follow, you are protecting an outcome the process was built for.

Pick one recorded call, thread, or session from the last month where you remember going off-script, or the one where you suspect you did.

If you have a checklist, template, or SOP you normally follow, paste it too. If you do not, the prompt will work from the routine visible in the rest of the same recording.

Take out client names, but keep your own so the model knows which speaker is you.

Paste the prompt below into Claude or ChatGPT, then paste the work underneath it.

🧠 Join CF Lab

The Prompt

Join CF Lab

Part 3 of The Judgment Economy

For about twenty years, scientists in the West couldn’t repeat a set of Soviet measurements.

The subject was a kind of high-quality sapphire crystal. The Soviet team had published their results. Western labs understood the theory behind it. They built or bought the same kind of equipment.

The results still wouldn’t match.

If you’re picturing a story about smart versus not-smart, that’s the wrong story. These were skilled scientists. They had the papers. They had the machines. They had everything that had ever been written down.

Progress came only when people showed up in person.

Researchers had to visit the Soviet labs. They had to watch the work happen with their own eyes. They had to see how the equipment behaved, which small adjustments the experienced technicians made, and what happened when the plan met the real material.

The missing piece wasn’t on paper. It lived inside the work itself.

🧠 Join CF Lab

What AI made cheap and easy

AI has made it much easier to produce a decent first draft of almost anything.

• Working code

• A readable article

• A working spreadsheet

• A research summary

• A passable slide deck

• A reasonable answer to a common question

A team of economists has been studying what happens when customer-support workers get AI help.. The biggest gains went to the least experienced workers. The tool spread out pieces of skilled performance that used to belong to only a few people.

That sounds like a productivity story but that’s the wrong headline.

When more people can produce work that’s good enough, “good enough” stops proving that someone knows what they’re doing.

The real story is about what gets scarce. Whenever something gets easy and cheap to make, value moves. It moves away from the thing that was hard to make, toward whatever decides if it’s worth making at all.

Making things is getting easy.

Deciding what’s worth making, and what’s actually right, is getting rare.

The Judgment Economy

We’re stepping into a world where information is everywhere, drafts are instant, and the thing that decides what happens next is judgment.

Judgment is noticing the one detail that changes everything.

It’s telling a real signal apart from normal noise.

It’s seeing the second problem before it shows up.

It’s knowing when the usual rule doesn’t fit this time.

It’s picking one answer out of several that all sound fine.

It’s deciding that a “good enough” draft still shouldn’t go out the door.

calls this “The Judgment Age” and says judgment is the scarce resource in a world full of easy output. He’s right about what’s happening, and I’m borrowing his diagnosis.

An age names the period we’re living through.

An economy asks who owns the scarce thing, who can pass it on, and who gets paid for having it.

If judgment is the scarce resource, the people who can show their judgment, hand it off, and correct it are about to matter more than almost anyone else in the room. The people who can’t will watch their hard-earned skill get buried under a pile of decent AI drafts.

Where You Already Feel This

This isn't a guess about the future. You already feel it in four places.

1. AI hands you options. You supply the standard.

The model can hand back twenty headlines, three strategies, or six possible diagnoses.

Someone still has to pick which one to use.

A beginner usually picks whatever sounds the best (AI can make anything sound good).

Someone experienced spends their time cutting, rejecting, narrowing, and adding context instead.

The beginner assumes the right answer is buried somewhere in the pile.

The person with judgment knows the pile is only a starting point.

Those cuts and corrections are where the judgment lives. That’s the real gap between having a model and knowing how to use one well.

2. Teams run on process. Exceptions land back on the expert.

A written process, an SOP, works great for the cases you've seen before, the ones that happen all the time. Blair Enns has a line for this.

Low variation in process gives you low variation in outcomes.

There’s typically no SOPs for the outliers and edge cases.

They keep climbing until they reach whoever has seen enough strange cases to know what the pattern means, often the founder or the senior partner whose judgment never got written down anywhere.

This is why you can hire good people and still end up being the bottleneck. The process was built for the normal case, and the normal case never needed you in the first place.

3. More content ships. Trust follows whoever can tell the good from the average.

Publishing more doesn’t prove you have anything worth saying anymore.

Anyone can find information now. Readers stick with people who keep spotting the problems worth solving, drawing the distinctions that matter, and seeing what happens next before the rest of us do.

The signal was never in the volume. It’s in what someone chooses to say and what they leave out. Post without that filter, and you’re just adding to the pile everyone else is already making.

4. Services get easy to copy.

A competitor can copy what you deliver. They can study your proposal, borrow your process, buy the same software, and end up with something that looks close enough.

Their weakness shows up the moment something changes…when something happens that wasn’t on the YouTube tutorial. Someone with real judgment knows which parts of a plan can bend and which ones hold the whole thing up. The copycat only sees a template.

That difference is judgment.

And it’s the one part nobody ever bothers to write down. It’s also the real reason a client should pay you instead of the cheaper copy.

Judgment isn't just a fancy word for opinion

You might be thinking, isn’t “judgment” just a nicer word for having an opinion?

No, because real judgment gives you the same answer on the same kind of case, again and again.

James Shanteau spent his career studying what actually separates an expert from someone who’s just confident. He found that experts who tell similar cases apart, and land on similar answers when a similar case comes back around, are doing something real.

That consistency is the proof.

Gary Klein studied how people make decisions under pressure. Firefighters, military commanders, ER nurses. He found they almost never sit down and weigh their options. They recognize a pattern. They’ve seen enough cases to know what usually comes next, and the recognition happens fast because the experience runs deep.

Judgment gets built through years of seeing things happen, feeling the consequences, and getting corrected.

That’s what gives it weight. It also means judgment leaves tracks…

The decisions made, the corrections taken, the moments someone left the plan because something felt off, the cases they still remember years later, the distinctions they can’t stop noticing.

Judgment isn’t some mystery. Most of it just never got written down.

Can AI actually learn judgment?

Yes, from whatever’s already been made visible.

AI can learn from stated preferences, recorded decisions, corrections, results, and examples. It’ll keep getting better at applying the standards someone bothered to write down. It’ll keep getting better at spotting patterns that show up in the data.

But everything an expert never wrote down stays invisible to it.

A model can’t learn from a correction that only happened in someone’s head. It can’t study the option someone rejected before anyone else saw it. Or sit in on the conversation where an experienced person felt the ground shift and changed course. It can’t pick up a standard that got applied but never got named.

The space between what an expert knows and what they’ve ever put into words is the same space between a decent AI and a genuinely good one.

Writing judgment down gives AI something better to learn from. Skip that step, and the model learns from the average of what’s visible, while the best parts of someone’s expertise stay locked away.

Back to the sapphire labs.

The scientists had the paper. What they needed was access to everything happening around it. The visiting researchers never turned into copies of the original technicians. They just got close enough to reproduce the result.

AI has made the same problem bigger. It can soak up almost everything visible in the world. The best parts of someone’s expertise still live inside decisions nobody recorded, like the quiet correction, the option turned down before anyone saw it, or the standard that got applied without ever getting named.

The next question is how to get those decisions down on paper without sanding them into some fancy five-step framework.

That’s what the next piece is about. But first…

Try it on one of yours

Every time you reject an AI draft, correct a team member, or rewrite your own first attempt, you apply a standard you have probably never written down. The rejections are the evidence.

Pick one recent AI chat where you pushed back more than once, or one draft you rewrote where you still have both versions. Take out client names, but keep your own so the model knows which turns are yours. Paste the prompt below into Claude or ChatGPT, then paste the session or the two drafts underneath it.

🧠 Join CF Lab

The Rejected Option Log

Cognitive Fingerprint Lab is $20 a month right now. On Saturday, September 5th, new memberships move to $25. Anyone in before then keeps $20 for as long as their subscription stays active.

Join CF Lab

Part 2 of The Judgment Economy

I was showing her what came out of a clinical session she had dismissed.

“I thought it was pretty boring,” she said. “And I wasn’t very helpful.”

The analysis had surfaced several pieces of content, recurring patterns, and decisions she had been using without ever naming them. She had seen nothing unusual because nothing unusual had happened from her perspective. Decades had been spent developing the ability to make those decisions.

Later, she said something else.

“This is 35 years of thinking from me.”

Thirty-five years of thinking had become a boring afternoon.

🧠 Join CF Lab

If you’ve ever solved something quickly and felt like you skipped a step, you know the feeling.

• A founder reviews an AI draft. The facts are correct. The piece still feels wrong. They rewrite half of it and can’t explain what standard the first version violated.

• A consultant joins a call after the team has spent an hour circling the problem. Two questions. The issue becomes obvious.

• A creative director rejects seven concepts in five minutes. Asked why, they point to details that were never in the brief.

What looks like a secret method is years of experience compressed into fast recognition. The result feels suspiciously easy. And when something feels easy, it’s hard to believe it’s worth much.

The Expertise Paradox

As someone gains experience, they stop consciously walking through every step. They start recognizing patterns and responding to the situation as a whole.

A beginner sees fifteen separate clues.

The expert sees one familiar shape.

This saves enormous mental effort. It also creates a strange side effect. The expert loses conscious access to parts of the reasoning that produced the answer.

This is the Expertise Paradox.

The better you get, the more automatic your judgment becomes. The more automatic it becomes, the harder it is to explain. The harder it is to explain, the less valuable the work starts to feel.

Michael Polanyi spent much of his career on this. His phrase was “we know more than we can tell.” Hubert Dreyfus mapped how experts move past rules and frameworks into something closer to recognition.

A chess master recognizes the position before calculating a single line.
Gary Klein found the same thing in fireground commanders. They sized up a fire and acted, with no checklist in sight.

The knowledge is all there. It has become invisible to the person who holds it.

Where This Shows Up

This isn’t just a therapist’s problem. I’ve watched it appear across four different domains.

1. AI keeps producing the average answer

A marketer I work with started using AI tools and ran straight into a wall.

She could tell the output felt like “telegram” when she wanted “magazine.” She knew the difference. She couldn’t explain what she wanted changed.

The model had access to every word but she held the standard.

Her frustration came from trying to communicate a judgment she had never needed to put into words. When you can produce a competent draft in seconds, the bottleneck moves from production to discernment.

The question stops being “can I make this?” and starts being “is this right?”

Answering that question requires standards most experts have never written down.

2. Your team needs you in the room

A founder I work with in the medical space has a team that can do the work. The hard part is transferring the patterns he uses in client conversations.

He was the bottleneck. Every unusual decision and every exception to the playbook climbed back to him.

Once I mapped those patterns, his team could walk into calls “super confident, not really worried if I’m on the call or not.”

The founder becomes the place where judgment goes when it can’t be transferred.

3. Your content describes the industry instead of your thinking

Another founder’s financial company has served its clients for years. Its published content was generic SEO material.

He put it bluntly.

“It’s not even anything extracted from the work we do. It’s all built for SEO.”

Anyone could find what the company published. Its most valuable thinking stayed trapped inside client conversations, in the moments when someone on the team knew something the industry didn’t.

4. You discount work that feels easy

The therapist had 35 years of experience and still treated an unusually rich session as unremarkable.

Another expert I work with put it this way.

“I work faster than my ability to tell you what I just did.”

Someone else sees speed and fluency. The expert feels like they skipped something.

That’s where underpricing begins. Doubt creeps in, and someone who should be charging for their judgment starts charging for their time, because time is the one thing they can prove they spent.

Why Introspection Usually Fails

So why can’t the expert sit down and write out their method?

Because when experts explain their work, they list what they remember learning…

The steps and the frameworks they can already name.

Those have value. They rarely contain the moments where expertise changes the result.

The missing judgment shows up while they’re doing the work. It’s in what they notice, what they correct, which rule they ignore, and what they refuse to accept.

The therapist couldn’t find her patterns by thinking harder about therapy. They were sitting inside a session she had stopped thinking about. The value had been there while she was living through it. She needed distance…to step back from her own work to see it.

The sessions experts dismiss are the first place to look.

Once I saw this happening across different experts, a larger question started bothering me. What happens when competent work becomes cheap enough to generate on demand? When AI can produce the draft, the proposal, and the code, the only thing that determines whether any of it is good is the judgment of the person looking at it.

That’s what the next article is about. But first…

Try it on one of yours

Pick the most ordinary piece of recorded work you have from the last two weeks. A call transcript. A client email thread where you answered a question. A Slack thread. A draft you rewrote. The one you would call boring.

You can take out client names, but keep your own so the model knows which speaker is you. Paste the prompt below into Claude or ChatGPT, then paste the work underneath it.

🧠 Join CF Lab

The Boring Session Test

Cognitive Fingerprint Lab is $20 a month right now. On Saturday, September 5th, new memberships move to $25. Anyone in before then keeps $20 for as long as their subscription stays active.

Join CF Lab

Part 1 of The Judgment Economy

Your Best Work Feels Suspiciously Easy. That’s The Problem.

A client of mine had 35 years of experience. She thought her session she recorded was “pretty boring.”

Then I showed her what actually came out of it.

Multiple content assets. Recurring patterns she’d been using for decades without naming them.

Judgment calls that took her two seconds, but would’ve taken someone else two years to develop.

She stared at the screen for a second, then said something that made me smile:

“This is 35 years of thinking from me.”

Thirty-five years of thinking...

Had become a boring afternoon. And this is not uncommon:

• A founder reviews an AI draft. The information is correct. But something feels... wrong. She rewrites half of it. Can’t explain why. It just was.

• A consultant joins a call after the team spent HOURS debating a solution. Two questions and only four minutes. He immediately knows what the problem is.

• A creative director rejects seven concepts in immediately. Points to details that weren’t in the original brief. Details no one else even noticed.

You’ve done this too. You’ve looked at something and instantly KNEW it was off.

But when someone asked you WHY, you had no sentence to explain it.

Here’s what’s happening. You’re not hiding some secret method. Your brain has compressed YEARS of experience into split-second pattern recognition. The answers feel suspiciously easy.

And when something feels easy, it’s hard to believe it’s worth much. So I started digging into why this happens.

Turns out... there’s a name for it.

Subscribe now

The Expertise Paradox

There’s a name for what’s happening here. As you gain experience something weird happens in your brain. You stop consciously walking through every step.

You start recognizing patterns. Grouping details together. Responding to situations as a WHOLE instead of individual pieces.

• A beginner sees fifteen separate clues

• The expert sees one familiar shape

This is evolutionary, it’s your brain’s way of conserving energy. Of not burning through mental calories analyzing shit you’ve already figured out. But it creates a strange side effect.

The expert loses conscious access to the reasoning that produced the answer.

Read that again.

The BETTER you get, the more automatic your judgment becomes. The more automatic it becomes, the harder it is to explain.

The harder it is to explain, the less valuable it FEELS.

Even though it’s actually MORE valuable than ever.

This is the Expertise Paradox.

Why It Matters Now

AI can produce a very good, confident-sounding first draft of damn near anything.

• A article

• A spreadsheet

• A presentation

• A piece of code

• A research summary

• A reasonable answer to most questions

These are things people built entire careers on. Things that used to separate the experts from the amateurs.

And now?

They’re commodities on a digital assembly line. Don’t believe me?

Try this...

Put out a feeler on LinkedIn asking if anyone can build you a new website. Watch what happens to your DMs. That site that used to cost 10K?

Now it’s $200.

Now ask all those people to show you PROOF their websites convert. Real numbers with actual results.

I’d be SHOCKED if 1 out of 20 can give you anything beyond vibes and Lovable heart bubbles. Here’s what’s happening.

• Execution is becoming a commodity

• Judgment is becoming scarce

And the economic pressure on that gap? About to get intense. When more people can produce “acceptable” work, the person who can tell you what quality actually looks like becomes the most valuable person in the room.

But here’s the problem…

Most experts can’t explain HOW they know. Their judgment has become invisible. even to themselves. And when you can’t articulate how you know something, it’s pretty tough to promote it.

Market it. Or sell it. But here’s the good news...

The same AI that commoditized execution?

It holds the key to unlocking the very thing that makes you irreplaceable.

What’s Coming

I’m publishing three articles on this. They’re designed to be read in order, but each one works on its own.

1. “She Thought the Session Was Boring”
   Opens inside the session my client dismissed. Explains why expertise becomes invisible to the person who has it. You’ll finish this one replaying moments when your own best work felt ordinary.

2. “Welcome to the Judgment Economy”
   Opens with a case where published information wasn’t enough. Scientists had the papers. They had the equipment. They still couldn’t reproduce the result until they SAW the work performed. Explains why AI makes invisible expertise economically urgent.

3. “Capture It Before It Disappears”
   Opens with a waste-management owner who could see a business inside a dumpster. Explains where judgment leaves evidence and what to do before it disappears.

The Judgment Economy

These articles are the start of something bigger. A change to what this newsletter is.

Since March of 2025, I’ve been building a methodology called Cognitive Fingerprint™.

The premise is simple.

Your best judgment leaves evidence in your recorded work.

• Your patterns

• Your blind spots

• Your mental models

• Your frameworks

(We all have them even if we’ve never named them.)

• Your decision DNA.

How, when, and WHY you make the calls you make. Basically, how you think.

The articles explore what that looks like from the inside.

• Where judgment leaves evidence

• How to capture it without inventing some polished fiction

• Why expertise becomes invisible to the person who developed it

• What captured judgment can become across AI, content, teams, and succession planning

Here’s the thing. What I’ve built over the past two years is too big to live in a newsletter. A newsletter can show you the territory.

I send once a week. You read it over coffee, maybe you nod along.

But the actual method. The thing that requires practice, distillation, and careful thought. That needs a place where you can bring your calls, your transcripts and run the tools against them.

Then you build your own intellectual property.

On Friday, September 4th, I’m opening Cognitive Fingerprint™ Lab.

A Skool community built around this method. Loaded with courses, prompts, installable Claude skills, recorded classes.

And a few surprises.

Inside, there are five pillars:

1. The Fingerprint. Why you can’t see your own expertise. And the science underneath it

2. The Map. The full territory sitting in one recorded conversation, charted. Most people run one extraction, get one result, and stop. We’re going deeper

3. Extraction. The prompts, workflows, and tools you run against your own calls

4. Your Method. Turning repeated patterns into a named method. A diagnostic. An SOP your team can run without you

5. The Guide Position. Pricing, positioning, and content built on what you found

Membership includes BOTH this newsletter and the Lab., no matter what door you come in through.

It’s currently $20 per month. On Saturday September 5th, new memberships move to $25 per month.

Current members and anyone who joins before the deadline keep the $20 rate as long as their subscription stays active.

Read the articles first.

If the series changes how you see your work, the CF Lab will be here. The first article publishes Wednesday.

My client thought her session was boring. Thirty-five years of thinking had become an ordinary afternoon. The articles that follow are an attempt to understand why that happens.

And what to do about it.

-Max

Subscribe now

Obsession is a dangerous trade. from Monsters of God

I found Monsters of God by accident on HBO. I was mindlessly scrolling for something to watch, but once I put it on, I didn’t get up.

I was hooked first by the characters, then the history, then the smuggling. It’s a crime story with every reptile I’ve never heard of in it. And the people…the people were exactly like you think they’d be. I kept thinking I’d turn it off after ten minutes and focus on work. I didn’t.

Hank Molt (great name) is one of the main characters and an apparent legend in that industry. He ran a pet store in Philadelphia and stocked it with animals he found overseas, often carrying them into the country in suitcases, which, customs barely checked.

Eric Goode, who created the series, says Molt liked that there were laws at all, because breaking them made it a better game. In the 1970s he went on what he called shopping trips with other dealers. The Wall Street Journal later quoted him on what they packed.

The gun to shoot someone, the machete to cut off our fingers if we were bit by a taipan.

There were no rules, a lot of private deals, and a snake that can kill you if you guessed wrong. They brought a blade to cut their own freakin’ fingers off if they were bit.

Then I got stuck on a price list.

Subscribe now

Names, Then Numbers

Molt circulated lists of rare reptiles, snakes and turtles and lizards with prices next to their names, and mailed them to collectors and to zoos that competed for the same animals. Before that, collectors and zoos bought and sold by calling each other, or by shaking hands.

But the list did something more important than advertise inventory.

People saw prices that seemed absurdly high compared with anything they’d seen before. Yet, rather than immediately destroying demand, the list helped make the animals feel more rare and more collectible. A reptile that had been an obscure creature in a distant part of the world became a named item in a catalog, available through a person who appeared to have unusual access.

→ The price not only reflected the value.

→ The price helped create it.

Then Washington made the sheet more powerful by accident. The Endangered Species Act came into effect in 1973 and Molt’s business boomed, because an endangered label meant rarer, and rarer meant he could charge more. A law written to protect the animals raised their price instead.

Other dealers noticed. Tommy Crutchfield, a Florida rival, ran Molt’s playbook, selling smuggled reptiles to collectors, pet owners, and zoos. Time’s reporting on the series puts his sales and trades at millions a year.

There’s No Pudding

Harry Beckwith named this in 1997 in Selling the Invisible, then said it blunter in 2022 in The Invisible Promise.

“There’s no pudding in a service, so there’s no proof.”

Beckwith said, “Your price is not just your price. It is a quality cue.”

When people can’t inspect the thing, they inspect the tag. He tells the old story about the carpenter who charged $2 for hitting the nail and $43 for knowing which nail. Charge for the years it took to know which nail.

My favorite piece of evidence in his book is about coffee. At $18, buyers of organic coffee got more passionate about it, and they could recite roughly twice as many details about what they were drinking. The higher number made people believe more, and study the thing harder.

The reptile list ran the same experiment with snakes.

Invisible expertise

The blank page test

The $43 in the carpenter story was for knowing which nail. That knowing is invisible expertise. It’s the judgment that only shows up when you’re in your zone of genius. When you’re operating from unconscious competence.

You listen for two minutes and you already know what’s wrong, and if someone asked you to write down how you knew, you’d stall, because the steps aren’t sitting there as steps anymore. They compacted. You got so good at the work that the work disappeared from your ability to articulate it.

Why you can’t say it

That’s why it happens. Repetition sank the skill below words. The better you are, the less you can narrate it. You’ve been doing it for 10+ years and the manual you could have written as you were building those patterns has disappeared.

What’s left is a feeling in the call, a tell in their voice, a thing you notice that a competent person with a checklist will miss.

The borrowed job title

When you can’t say it, you sell familiar language or you borrow a job title. Someone asks what you do. You say you coach, or you consult, or you help people with the thing you’ve spent 10+ years getting really good at. The sentence comes out different every time, because you honestly aren’t quite sure HOW you do what you do. They nod, then they ask what you charge and if you are how I used to be you say “it depends.”

You leave the conversation having been treated like a less expensive version of someone you’re not, and you can’t correct it because you still don’t have a sentence for the actual thing.

I used to do this on my own calls. I can sit with a client and name their offer in one pass, then get on a sales call for my work and describe it as a pile of verbs (help, advise, extract, train). The work was there, I just never put it on a sheet.

What I Put On Mine

One line on my list is called Cognitive Fingerprint™.

The person who buys it is a founder whose team still has to wait for them after a call, because the knowledge is in one head. When we’re done they have a written extraction they can hand to people… their patterns, framework…their thinking, in their language.

That’s the whole card… a name, who it’s for, what they walk away holding, who it’s not for, and a price, which I put on the card I send them, not in this essay.

Here it is, item 001 on my actual sheet.

I have other lines with the same idea. Each one is something a buyer can say yes or no to without me giving a tour of my whole brain.

I can’t tell you how many opportunties I lost at workshops, calls or conferences when people asked how I could help them. I knew I could but I couldn’t articulate it.

The Gap

You can see what the sheet did, and you saw one of mine. Now it’s time to write yours.

The Monster Price List is the prompt (or the Skill) that digs that sheet out of your own calls. You feed it one to three transcripts of you working (your website and LinkedIn help, but the calls are where the expertise actually shows up), and it hunts for the moments a buyer valued… the “how did you know that,” the thing you fixed for free, the “could you also” you never invoiced.

Then it writes one to five cards… a name pulled from words that were actually said, a buyer you can picture, something they walk away holding, who it’s not for, a number they can argue with. Every card is quote-locked to a real moment from your own tape. You shouldn’t have to sit through a questionnaire first.

The full prompt and the downloadable Skill are below. Attach a transcript (or 3) of a coaching call, discovery, onboarding etc.. Send the first card.

It works from the moments your expertise already leaves on tape. It won’t map the whole of how you think (that’s a deeper extraction), but it will name and price what you already sell right now.

Read more

And speaking of a good time, I spend way too much of it on X and YouTube, bookmarking all the updates that I will “read later.”

But my bad habits are your good fortune as I do come across lots of good Skills and other AI goodies.

Below are 5+ Skills, including a few collections for you to experiment with this week…

Let’s boogie.

Subscribe now

🛠️ The Opus 5 Skills Upgrade Prompt

Creator: Matt Shumer, Something Big Is Happening (somethingbig.ai)

Link: https://somethingbig.ai/skills-upgrade

Every skill you built for an older Claude model could be working against you now.

That’s the pitch, and it’s an “ah hell…” one if you’ve got a .claude/skills folder you’ve been adding to for months. Skills written for older models tend to over-explain and over-constrain, patching around weaknesses the newer model doesn’t actually have anymore.

Paste Shumer’s prompt into Claude Code, wherever your skills actually live, and it gets to work. First it finds every skill you’ve got installed and backs up an untouched copy of each one before touching anything. Then it builds a live HTML dashboard right on your own computer, laying out the whole inventory: what does what, which ones came from where, which ones you’re even allowed to touch, and where two skills are quietly stepping on each other. Then it goes through what he calls a Gauntlet Loop: its own builder, its own blind critic. Beat the original at its own job, it’s an upgrade. Lose, it gets retired.

The fun version: pick the ONE skill in your library you’re proudest of, the one you spent real hours tuning, and send it through the gauntlet alone first. Blind-judged against bare Opus 5, no instructions at all.

✏️ Human Review

Creator: , Behind the Craft (creatoreconomy.so)

Link: https://github.com/petergyang/human-review

You know the paragraph you end up typing just to get one typo fixed? List every change, hope the agent catches all of them, then go double-check its work anyway.

Peter Yang built the fix for that. If his name’s familiar, he’s the guy behind /no-ai-slop, the skill that hit 4,500-plus GitHub stars for hunting down AI writing tics like “Most creators…”, “writing that’s actually yours”, and the “it’s not x it’s y” that just WILL NOT go away.

Human Review tackles a different half of the same problem…Reviewing what AI builds, not just what it writes.

• Point it at an HTML or Markdown file, or a page running on localhost, and it opens it right in your browser with a feedback panel next to it.

• Edit the text and basic formatting yourself.

• Drag an image corner to resize it.

• Click any phrase, photo, or section and drop a comment directly on it, the way you’d comment inside a Google Doc.

• Command-click a link to jump to the next page in a multi-page site without losing what you already flagged.

• One click on “Send to agent” and it gets every edit and every comment in ONE batch, not five chat messages.

Nothing leaves your machine. No account, no API key, no cloud upload.

One quirk: HTML edits save the second you make them. Markdown and localhost pages need that Send click before your agent sees the change.

Use it any time you’re reviewing something an agent built instead of something it wrote: a landing page, a planning doc, a multi-page site.

Install by pasting “Install the /human-review skill globally from https://github.com/petergyang/human-review” into Claude Code, then run /human-review index.html or /human-review localhost:3000.

The fun version: leave the kind of blunt note you’d leave on a coworker’s draft. Click the stock photo and just type “no.” Click the paragraph that’s doing nothing and type “cut this.” Send it, and watch your agent act on the note instead of guessing what you meant.

Full Peter write-up here:

Behind the Craft

Use My /Human-Review Skill to Edit HTML and Markdown Files Visually

Dear subscribers…

Read more

a month ago · 49 likes · 3 comments · Peter Yang

📚 The Skill Shelf

Creator: 🙋🏻‍♂️ ←

Link: https://aistrategyexchange.com/skillshelf

I started a collection of ones I use all the time or send to clients/friends and tossed them on the shelf. They free. Enjoy.

🎨 Impeccable

Creator: Paul Bakaus (@pbakaus on X)

Link: https://impeccable.style/

I bet you recognize it even if you don’t know you do.

The cream background. Oversized italic serif headline. A soft glow sitting behind the hero.

These are design defaults every coding model can’t help but use over n over, and Paul Bakaus built the tool that stops it.

One skill, 23 commands, and it works across nearly every AI coding harness that exists. Claude Code, Cursor, GitHub Copilot, Codex, Gemini CLI, Grok Build, more. It does two jobs.

1. It steers what gets designed in the first place, reading your tokens and components so new work fits your system instead of overwriting it.

2. And it catches the tells if they slip through anyway, a 58-rule detector that can run in CI and fail a pull request, or scan any page in your browser, including a competitor’s.

V4 just shipped and Bakaus’s team ran an experiment, asking a coding model to be creative sixteen different ways. 30 of 35 answers came back as the IDENTICAL concept. So V4 stopped letting the model pick its own favorite direction. Now it deals a hand of curated design “worlds,” 188 of them, human-reviewed out of 360 candidates, as challengers against whatever the model would have defaulted to. You pick the winner yourself on a decision page that opens right in your browser.

The core prompt also got rebuilt leaner, on purpose, because newer models like Fable 5 follow short structured instructions better than long insistent prose (see: Shumer skill).

Use it any time an AI agent is building or redesigning your interface and you want it to look considered instead of templated. Install with npx impeccable install, which detects your harness on its own, or grab it through the Claude Code plugin marketplace. Run /impeccable init the first time.

The fun version. Start something low-stakes and let the dice pick your design direction instead of you picking one. The deck’s got a world built on W.E.B. Du Bois’s 1900 data portraits, another modeled on split-flap departure boards, another straight out of Saul Bass movie titles. Deal yourself something you’d never choose and see what happens.

📱 The Ultimate Claude Content Pack

Creator: , founder of Blotato

Link: https://help.blotato.com/claude-skills/claude-skills

Seven free skills, and they’re built to hand off to each other so you only ever have to talk to ONE of them.

Sabrina Ramonov built this and is the one behind Blotato, the tool that handles the posting, but she also built an NLP startup years before ChatGPT existed and sold it to Pega Systems. These days she personally ships 250-plus pieces of content a week (😯) off this exact stack, so the pack had to survive her daily use before it ever reached the public.

The chain kicks off with content-coach. Tell it you don’t know what to post and it checks for a brand brief, runs a five-question intake the first time only (what you sell, who buys, your CTA, a recent story, your vibe), then hands you ideas. Pick one and post-writer drafts the hook, body, and CTA. post-grader scores it, hook strength alone is half the score, since the first three words decide whether the post survives the scroll, and loops the draft until it clears an 8 out of 10. Approve, and post-scheduler ships it.

Two more skills sit on top for when you’ve got more than an idea.

The first, repurpose, turns one long piece, a blog post, newsletter, or transcript, into a week’s worth of platform-specific posts (three for LinkedIn, five X threads, two short video scripts), opening each one with a line pulled from viral-hooks’ library of 100 tested hooks.

You don’t need Blotato for any of this except the very last click (though I’ve heard good things). 6 of the 7 skills run on nothing but Claude and file access. post-scheduler is the only one that touches Blotato, and that one just writes the finished post to a file for you to paste yourself if you’re not connected.

Install the whole pack as one ZIP. Unzip it, drop the seven folders into ~/.claude/skills/, restart Claude Code or Desktop. On claude.ai or Cowork there’s no local skills folder, so you upload each SKILL.md one at a time through Customize > Skills.

The fun version. Hand repurpose your worst, most rambling old blog post and let it try to mine three LinkedIn posts and five tweet threads out of it anyway. Then let post-grader tell you, in a number, exactly how bad your hooks are compared to hers.

📬 P.S. — The Sneaky Useful Update

Link: https://code.claude.com/docs/en/cross-session-messaging

Claude Code sessions can now message each other. Working across a few terminals or worktrees, you can ask one session to pass a heads-up to another instead of retyping the same context yourself. Codex and Cursor have had a version of this for a while. Claude showing up late to the party is fine. Being there is what matters.

One catch: if your session runs with permissions bypassed, which a lot of ours do, Claude Code holds every incoming message for your approval by default. You end up clicking accept all day instead of getting anything done. Fix is one line in ~/.claude/settings.json:

{
  "crossSessionInbound": "accept"
}

Or just tell Claude to add it for you. Saw this fix floating around on X and it’s too good not to pass on.

Go get’m this week!

-Max

Zero crashed in the Akutan bog. US Navy / NARA.

The island sits in the Aleutians, a chain that looks, on a map, like someone spilled gravel between Russia and Alaska. Fog sits on the water for days. The ground that pretends to be a landing strip is often marsh.

Petty Officer Tadayoshi Koga, flying off the carrier Ryūjō during the raid on Dutch Harbor, put his Mitsubishi A6M down on what he thought was flat ground. The plane somersaulted and Koga died still belted in, hanging inverted in the wreck.

His wingmen circled and the standing order said to destroy a downed Zero so the enemy couldn’t study it. They didn’t fire because they hoped he might still be alive.

He wasn’t, but the plane mostly was.

US crew inspects the recovered airframe on Akutan Island. US Navy.

Subscribe now

That single intact airframe, later recovered by the U.S. Navy, shipped south, rebuilt, and flight-tested in San Diego, taught American pilots more about beating the Zero than months of briefings that described the same fighter from a distance.

The Zero loaded onto a ship for transport south. US Navy.

I keep thinking about that Zero when I sit with founders who already “know” what their employees are dealing with.

They have the reports. They don’t have the Zero.

What the briefings already said

By mid-1942, American pilots had met the Zero in the air and hated the experience.

• It climbed past them

• It turned in between them

• In a climb-and-turn fight, they kept losing

Intelligence even had a detailed file which reported light construction, exceptional low-speed agility, thin armor, no self-sealing tanks.

But here was the problem…

A briefing can tell you a plane is light. It cannot tell you what the stick feels like when you push past 200 knots.

In September 1942, Navy test pilot Eddie Sanders took the repaired Akutan Zero up for its first flight at Naval Air Station San Diego. He flew about two dozen tests scenarios over the next few weeks.

On the very first flight, American pilots saw and experienced, exposed weaknesses they could immediately use.

Eddie Sanders taxis the captured Zero at NAS San Diego. US Navy.

• The stick stiffened hard at high speed

• Rolling at that velocity was slow and created heavy force on the stick.

• The plane rolled left more easily than right

• Under negative G, the engine cut out because of the float-type carburetor

None of that transferred in the briefing like it did up in the air.

It’s stick feel. Asymmetry. An engine that quit when you asked for a specific kind of acceleration.

In mid-fight, pilots don’t get time to think about facts.

They have fractions of a second to react. A briefing in a folder doesn’t help when another plane is already on you. What helps is body knowledge you can fire without a meeting.

During The dive, open range while the Zero’s engine is coughing, and at about 200 knots roll hard right before he can settle his sights, because his high-speed roll is a slow brick and yours isn’t.

That’s what the intact plane bought.

The Zero in U.S. markings on a test hop over San Diego. US Navy.

The war already had piles of intel. It needed one plane you could put a test pilot in.

Your company already has Zeros

If you run a team that sells, coaches, installs, supports, or closes anything complex, you live on a version of this problem every week.

Buried in your processes is your Zero.

These are the moments hidden in plain sight. They live in coaching calls, brainstorms, sales calls…anywhere that a person is problem-solving and tapping into their expertise…their unconscious competence. Many times, we feel them when things don’t go according to plan but we keep the conversation going anyways.

And the craziest part, it often happens without knowing why…we “just knew”.

When we try to investigate we will review our slides. Webinar decks. Meeting agendas. CRM stages. But those are all briefing material.

What you almost never think to look at is the intact plane.

One moment preserved with enough detail that you can spot the seemingly invisible.

My favorite place to go hunting is the call logs. A transcript gives you more than a label.

• The sentence where the prospect’s energy dropped.

• The ten-second pause your best rep stayed quiet…and then the prospect answered their own objection.

• The implicit expectations that when you deliver on, makes you look like a mind-reader.

I work with businesses who are drowning in this frontline intelligence and still lose the same fight on a loop.

The knowledge exists. It lives in our har drives and in the cloud. They are full of “you had to be there” stories that stay hidden or forgotten unless you go in an extract them.

It’s time to find your Zero

You now know why America needed the Akutan Zero more than another folder of Zero reports. Field knowledge beats distant description when you can re-fly the machine. Your company is full of these moments that die in Slack threads and pipeline adjectives. One intact story, captured with handling detail, can teach more than a quarter of debriefs.

The Intact Capture is the prompt that turns one story (a sales call, a coaching session, a site visit, a support recovery, an ops near-miss) into a reusable knowledge object. It forces the scene, the handling secrets, the asymmetries, and the exact moment the engine cut. It writes a green-lie check (what your CRM or checklist will still mark as fine), a stop rule, and a “Do this tonight” block so the knowledge leaves the call and enters the team the same day.

Below the paywall is the full machine. Path A if you can interview the person who lived it. Path B if you only have a recording, transcript, or messy notes. Run it once this week on a real fight you’re still losing. Get the plane on the ground. Then let someone else fly the test.

How to run it

1. Pick one live fight… a deal, install, escalation, or renewal where your average process keeps losing.

2. Choose Path A (15-minute interview with the person who was there) or Path B (recording, transcript, or notes).

3. Paste the prompt. Fill the capture target fields with ugly honest detail. Run it.

4. Do the “Do this tonight” block the same day. Send the message. File the knowledge object. Try the one behavior on the next similar call.

5. After one real re-flight, answer the recovery questions and run the prompt once more. Keep the second version.

Read more

The bed is made wrong. A calendar on the wall is frozen on the month of the incident. The later pages are still attached. Lee knew good detectives clock time clues like that, including what is not advancing.

A window opens. The fabric on a chair is worn the way fabric gets worn when a real person sat there for years. The scale is one inch to one foot.

Everything is tiny and nothing is typical dollhouse-cute.

If you stood over this table in the 1940s, someone would start a stopwatch at ninety minutes and tell you to study the whole room. Start in one corner. Work every wall, every object, every surface. Then do it again through the middle. Not a circle.

Just watch the room and notice what “feels” off.

The woman who built these rooms was Frances Glessner Lee.

She was a Chicago heiress, born in 1878, denied from the medical path she dreamed of. Instead, she used her inheritance to fund a department of legal medicine at Harvard Medical School in 1936. Then she did something stranger than putting her name on a building.

She built dollhouse dioramas of death.

She called them the “Nutshell Studies of Unexplained Death”, after an old police line about finding the truth in a nutshell. There were about twenty of them. Built with a carpenter over years, in wartime material shortages, some running what would be six figures in today’s money. She even went to autopsies so the bodies would look right.

She wore out-of-date clothes herself so the fabric would age the way old clothes age. She knitted tiny victim clothes under a magnifying glass, a few minutes at a time before her eyes quit.

The rooms were composites of real cases, not fantasy.

• Sex workers

• Messy kitchens

• Domestic violence

Spaces nothing like the houses she grew up in (I hope).

In 1946 she started handing them to investigators for training, as part of a larger system.

Seminars that put cops and medical people in the same room, week after week. Harvard later dissolved the department for internal political reasons and the dioramas went to the Maryland medical examiner’s office in Baltimore. Today, they are still used in forensic seminars.

And in all the weirdness of dollhouse murder scenes and heiress forensics, that last part is the piece that actually matters for this article.

Subscribe now

Optional Watch: The Full Frances Glessner Lee Story

If you want the full Frances Glessner Lee story (the Harvard seminars, the wartime workshop, the rocking chair that had to rock the same number of times as the real one), this documentary is worth the watch:

These nutshells were a training environment for judgment, built so carefully that detectives would trust them, and kept alive by a teaching system that outlasted any single institution.

Why a manual still fails (and a “room” works)

Lee could have written a manual.

She had the money, the access, the medical connections, and the time. Best practices would have told her a manual would have scaled on paper.

But it also would have sat on shelves. It would have been photocopied into oblivion and ignored by every new detective who still walked into a real room and missed all the important clues.

So, she built rooms instead because she knew a manual transfers vocabulary. A room transfers seeing.

The point of the Nutshells was not to “solve” the little dollhouse crime and high-five. It was to force investigators to process evidence so they did not make a hasty call. Officers later said they got better at real scenes because they had been forced to think their way through these ones (a novel idea, I know).

You can memorize the chapter on blood spatter and still miss the calendar pages missing.

You can pass the quiz and still walk past the worn chair.

The full-room-focus forces you to stay in the environment long enough for your judgment to wake up. The ninety minutes are a container for attention.

I watch something similar in businesses every week.

Someone on the team tries to make a decision that only the founder could make.

So they write the SOP with clear steps, share it in Slack. Maybe a Loom. For a day or so, it feels like progress and then a real situation shows up, and the team pings the founder again because they just don’t feel confident in their answer.

The document exists. The judgment does not.

Or the more common, 2026 version. The company buys AI tools where everyone has an overly eager digital assistant and a chat window. But when the employees screen share for five minutes, you see people open a blank conversation and start from zero.

To make matters worse, you see them quickly take AI’s first response as gospel and move on with their day.

That’s a modern “Nutshell Study of Unexplained Context”. The scene is missing and the assistant is guessing.

Two Failures Modes That Stack

Time and time again, when the team’s are producing slop, these are the two biggest culprits:

• Culprit 1: People take the first answer from AI. They skip the sit. Lee forced ninety minutes of looking before anyone got to declare a theory. While it may look like busywork, it’s actually building the muscle of thinking the problem through.

• Culprit 2: Even a careful sit fails if the room is empty. No transcripts, no judgment file and no instructions that say “run the pass before you publish.” So you get a clever rewrite of the wrong problem.

One of Lee’s teaching points, still used in Baltimore, is:

Don’t let what you already “know” keep you from what you should know.

What a business nutshell looks like in real life

Here is what I mean by a nutshell in a business, without the blood and the doll furniture.

Last week I sat with a founder who wanted his team to own their numbers. Not “know the KPI.” Own it. Know what to do when they were behind. Know when to escalate and when to solve it themselves.

• He had folders

• He had projects

• He had AI access

On paper he had already built the house (hell, it was a freakin mansion).

When we looked at how people actually worked, almost nobody was living in the house. They were standing on the porch every morning with a new blank chat, inventing the day from scratch. The knowledge never compounded because there was no scene. There was only a tool and some hope.

So we did one small experiment. We took a single recurring decision, the kind that always bounced back to him, and we built a Nutshell for it.

The decision was ordinary. When is a client proposal “ready enough” to send, versus still living in founder polish mode.

If you write that as an SOP, you get something like:

1. Check pricing

2. Check scope

3. Have sales review

4. Send

Directionally right but operationally…generic.

A new hire can follow every step and still ship something that fails the founder’s real test, because the real test is not on the list.

• The real test is a feeling in the stomach when a paragraph is still performing instead of promising

• The real test is knowing which missing line item will explode in week three

• The real test is a pattern the founder can spot in eight seconds and cannot yet put in a bullet

So we built a scene instead.

🕵🏻‍♂️ The Scene: One Proposal, Client Eagerly Waiting

This is the “room” for that decision. Imagine you are looking over the founder’s shoulder.

What is on the desk / screen:

• Google Doc open: Proposal_v4_FINAL_use-this-one

• Three browser tabs (client site, competitor PDF, HubSpot)

• Slack open: client message says “any update?”

• Printed competitor one-pager on the desk

• Sticky note on the monitor: “don’t underprice again”.

That is all surface inventory. Next is what a careful look finds.

🔎 Clues A Smart Person Walks Past

Same objects with a deeper read. This is what the SOP never captures:

• The doc title says FINAL. Version history says otherwise: the “why us” paragraph was rewritten four times. A smart person polishes the adjectives. The clue is that the proof is still weak.

• The competitor page is only marked on the results section. Not the services list. A smart person copies more bullets. The clue is buyers are buying certainty, not scope.

• Slack is polite. “Any update?” A smart person rushes the send. The real heat is in an email from two days earlier that is not even on this screen.

• The sticky note looks like a pricing reminder. It is not. It is about a discount pattern that taught the wrong lesson last quarter. A smart person lowers the price. The clue is a different kind of fear.

You can follow a checklist and still miss every one of those.

🤖 Where AI Typically Goes Wrong

It rewrites the proposal to sound sharper. It adds buzzwords and “improves clarity.” It does not notice the founder is stuck on proof because the last three deals died when buyers could not see themselves in the case study.

📑 The Full Room Pass (How To Study The Scene)

This is Culprit 1 in practice: Sit before you decide. Do not take the first theory. Do not take the first AI paragraph.

Historians and trainers often describe Lee’s eye-path as a spiral. Think of it as a full room pass.

Start top-left. Sweep the walls, the furniture, the floor, the open door. Hit every surface. Then come through the middle again. Stay in the room until you stop inventing a theory and start noticing what is actually there. Up to ninety minutes. No jumping to “I already know what this is.” No opening Claude until the pass is written down.

In the business version, the pass is the same job with different furniture. Before anyone opens a tool or types a prompt, they walk the scene in a fixed order for about twenty minutes:

1. What is this clients history?

2. What was the last communication?

3. What are the second-order effects of this decision?

4. Where are my blindspots? Where are theirs?

5. What is the first question they will ask when they read this and can we answer it in advance?

Write what you see before you write what you would do.

I ran one employee through that scene without giving the “right” answer. Their first instinct was to polish the language. After the pass, they asked a different question.

“Do we have one proof story that matches this buyer’s exact fear, or are we still selling services?”

That is the sentence the founder usually has to produce alone.

One scene. One decision. One person who got a little closer to seeing what he sees.

When this works with AI, it is not because the model got a clever one-off prompt. It is because the context was already connected.

• Past transcripts in the same place

• Knowledge files actually current

• Custom instructions written so the model can catch the nuance in the situation and answer strategically, instead of sounding good while missing the real problem

That is lesson 1 and lesson 2 working together.

You sit with the scene first (human or AI, same rule: no first-draft answer as final). And the system is already loaded so the sit has something real to stand on. The room is not only for the human. The room is what you install so the right context shows up at the right moment.

Half-built context fails the same way a sloppy Nutshell would have failed. If the details are wrong or thin, people (and models) stop trusting the room and go back to guessing.

The Prompt: You Understand It, It’s Time To Install It

1. Do not take the first answer. Sit with the problem. Build the muscle of thinking it through while the scene is still in front of you. Lee’s ninety minutes were that muscle. Your full room pass is that muscle. Hitting enter on Claude’s first draft is the opposite.

2. Have the context and the system that support that sit. Transcripts, judgment files, custom instructions, triggers that surface the right room at the right moment. Without that, you are either guessing alone or polishing the wrong problem with a smart model.

Frances Glessner Lee built dollhouses so people could practice lesson 1 inside a trustworthy room. She ran seminars so the room lived in a system, not a drawer. Your SOPs fail because they skip both.

The Context Nutshell is the productized prompt that turns one recurring decision into a scene, a clue inventory, the traps a novice (or a blank chat) will fall into, and install instructions so the context can live in files, custom instructions, and the moments that should trigger it.

If you already have an SOP, you can attach it. The prompt will use it as the old manual and build the room and install pack around what the SOP misses.

The point is to make the missing scene legible, and to wire it where people and AI actually work.

Below is the full prompt. Build the Nutshell. Install the context. Then the next time the same problem shows up, force the sit before the first answer.

🥜 The Context Nutshell

Copy everything below into a fresh chat with your AI of choice. Replace the bracketed sections with your real situation.

Optional but powerful: Attach or paste your current SOP (or checklist / Loom transcript) in the SOP block.

The tool works with a blank start or with an existing manual. Do not paste your “correct answer” or the founder’s preferred choice into the prompt. If you do, you are writing a manual again.

Prompt:

Read more

And sometimes they cry.

Not sad crying. The other kind. The kind that shows up when something true finally hits home.

For a long time I didn’t know why so I went and read what scientists have figured out about this exact feeling.

Here’s what I discovered, explained in a simple way, because it turns out this is one of the most interesting things about being a person.

This has a name, and it’s older than me

I thought I’d stumbled onto something brand new. I hadn’t. Scientists have been studying “what happens when someone shows you a true picture of yourself” for years.

They just used bigger words for it.

Turns out, there’s a kind of therapy where the whole point is to hand someone the results of a test about themselves and talk it through.

Here’s what they found when they added up all the studies. The biggest change wasn’t in fixing problems, though that mattered plenty. It was in the person feeling understood.

Feeling seen.

There’s also an exercise where people ask their friends and family to tell stories about them at their best. Then they read all the stories together. People come out of it lit up, full of fuzzy feelings, ready to go take on the world.

So the thing I do has cousins. And the cousins all point to the same clue.

Being shown a true, kind version of yourself does something powerful to a human being.

Subscribe now

Why the tears? You have two brains, and they disagree

Scientists figured out that when someone gives you feedback, two different parts of you react, and they don’t always agree.

One part is your thinking brain. It asks:

“Is this true? Does this match what I already believe about myself?”

The other part is your feeling brain. It just asks:

“Does this feel good or bad?”

Now picture a talented person who secretly thinks they’re not that good. This happens all the time, even to experts. Maybe especially to experts.

You show them proof of how brilliant they are. What happens?

Their feeling brain says yes. Right away. Warm. Happy. Their feeling brain never argued with good news in the first place.

But their thinking brain has spent years saying “I’m not really that good.” It wants to push the compliment away. It wants to say “you’re just being nice.”

Here’s the trick, and it’s the whole secret of my work.

When the proof is their own words, their thinking brain runs out of arguments.

It can’t say “you’re just being nice,” because the words aren’t mine. They’re theirs.

Caught in the act of being great.

So the thinking brain finally stops fighting. And when it stops fighting, the feeling that was there the whole time comes rushing up.

That’s the crying. That’s the sound of the fight ending, of a person finally letting themselves believe something good they could never believe before.

This is why proof beats praise

You might think, “Why not just tell people they’re great? Why the whole map?”

Because praise and proof aren’t the same thing, and your brain treats them completely differently.

Scientists found something almost spooky here. If you praise someone way past what they believe about themselves, it can actually make them feel worse. Their body treats it like a threat. It doesn’t land like a gift. It lands like pressure, like a lie they now have to live up to.

Proof is different.

Proof doesn’t ask you to believe a nice story. Proof shows you the thing you already did. You can’t feel like a fake about something you can watch yourself doing.

So the rule is simple. Don’t flatter people. Show them the receipts.

What happens after the tears

The crying is just the doorway. What matters is what walks through it.

When people can finally see and name what they’re good at, they start doing things differently:

• They raise their hand more. When you’ve got a clear, honest picture of what you bring, you stop shrinking. You say yes to the room you used to hide from.

• They make better choices. People with a clear sense of their own strengths make sharper decisions about what work to take and what to charge for it. A fuzzy sense of self makes fuzzy choices. A clear one makes clean ones.

• They can finally hand things off. You can’t teach or delegate a skill you can’t name. The moment it has a name, you can give it to someone else. You stop being the only one who can do the thing.

Naming your genius does more than feel good. Once you can name it, you start using it on purpose.

The twist that makes me careful

There’s an old story about a centipede.

Somebody asks the centipede how it walks with all those legs, which one moves first, how does it keep them straight.

The centipede starts thinking about it. And then it can’t walk at all. It trips over its own legs. It had never needed to think about it, and thinking about it broke it.

That’s a real thing, and scientists have watched it happen.

Take an expert golfer. Their body knows how to putt. It’s smooth and automatic (“it’s all in the hips”). Now make them think hard about exactly how their arms move while they do it. Their putting gets worse. The thinking gets in the way of the doing.

When I play tennis the same thing happens to me. The more I think about my serve, the more likely I am to hit my partner in the back of the head.

But here’s the weird part of the weird way our brain works.

This only happens to certain kinds of skills. The smooth, automatic, in-the-moment ones, like putting a ball or hitting a serve. Those can break when you shine a light on them.

The other kinds of skills are safe, and even get better when you name them.

• How you make a judgment call

• What you choose to focus on

• How you frame a problem

Naming those doesn’t break them. It hands you a new dial you didn’t know you had.

So there’s a line.

On one side, naming a skill makes it stronger.

On the other side, naming it can trip the centipede.

Here’s the good news. Even the tricky skills can be protected. Get someone used to being watched, little by little, and the tripping fades. It’s not a wall. It’s something you can train your way around.

But it means my job isn’t just to show people everything. It’s to know what is safe to shine a light on, and what is better left to run in the dark, doing its quiet, brilliant thing.

And one more honest thing

I won’t pretend the science answers everything. It doesn’t, and I’d rather tell you that than sound more sure than I am.

We know that naming your feelings helps you handle them. Whether naming your skills works the exact same way, nobody has actually tested yet. It’s a great guess. It’s not a proven fact.

We also don’t yet know how long the change lasts. The tears are real. The good feeling in the room is real. Whether it’s still working a month later, when you’re back in your normal life, is something no one has measured well.

Those aren’t holes to hide. They’re the next questions worth chasing. And honestly, almost nobody is chasing them, which means there’s room to be the one who finally answers them.

If you remember one thing

Someone shows you a true picture of your own genius, built from your own words.

Your feeling brain says yes right away.

Your thinking brain finally runs out of reasons to say no. The gap between them closes, and that closing is the tears.

Then you carry a clearer version of yourself back into your life. You raise your hand more. You choose better.

You can finally teach what you do.

And a wise guide knows there’s one thing left to protect.

The handful of skills that work best when you let them run without watching.

Even a genius should leave a few of their legs alone.

The part you can’t do by just understanding this

You’ve got the mechanism now. A true picture works because it gives your thinking brain proof and your feeling brain permission at the same time.

But knowing that and being able to do it on yourself are two different jobs.

Left alone, you’ll either flatter yourself, argue with the compliment until it dies, or shine a floodlight on a skill that worked better in the dark.

I built The Proof Mirror for that gap. It’s a prompt or Swkill workflow you run on your own material, not your self-opinion. It pulls patterns from call transcripts, emails, notes, and decisions, separates proof from praise, and flags the handful of skills you should name gently or leave alone.

The setup checklist, the extraction prompt, the proof audit, and the centipede check are below.

Read more

We were talking about agents, Claude Code, workflows, install paths, new tools. All the shiny stuff that makes everyone feel like the future is either already here or three terminal commands away.

Then Dionne (a member) said something that made me smile:

“I don’t want more input.”

Yes. Yes. Yes and yes.

People talk about agents like a coworker that runs on its own. Something that thinks, acts, hands off work, files things, writes drafts, keeps watch, sends updates, remembers it all, and politely asks you to approve something near the end.

It sounds great in theory.

It also sounds exhausting if your life already has too many tabs open.

Think about it. Agents get sold as the thing that takes everything off your plate. But if your new digital buddy hands you another inbox, another brief, another homework assignment, another pile of stuff to check before you can start your real work...

Then the agent is just one more thing to manage. One more place for work to pile up.

Subscribe now

The Wrong Starting Point

Most people start with the question:

“What agent should I build?”

That drops you straight into tool land or whoever is shouting the loudest about the latest.

You start comparing frameworks. You watch someone build a CRM assistant on YouTube. You install something globally, then locally, then inside the wrong folder, and then the terminal yells at you in what sounds like a haunted printer.

Eventually you get something running.

Then comes the question I hear from almost everyone after they install something like OpenClaw for the first time:

“What do I do with it?”

Then there’s the other side, which is where I usually live. I install too much.

I over-build. I assign, I create, I keep going until there’s so much running I’m not even sure what I just did.

My phone kept buzzing with Telegram updates. New content ideas. Status updates on clients. Reminders to do something I swear wasn’t on my calendar that morning.

And here’s the real problem. None of the things it kept pinging me about were things I actually needed to do.

Now my whole day went to checking those messages, chasing new rabbit holes, bouncing between things and never finishing any of them.

My new employee had just made a new job. For me.

The Secret Elf Test

So instead of building something that makes more noise all day, what if we asked a better question:

“What should already be done when I show up?”

Picture a store before it opens.

Someone turns on the lights. Someone counts the register. Someone sweeps the floor. Someone pulls yesterday’s notes. Someone sets the right thing in the right place so the real work can start clean.

Nobody calls that “running the business on its own.”

They call it opening.

That’s the secret elf version of an agent.

Open the store before I get there. Do the small, quiet prep before I show up. Set the right thing on the counter.

For a founder, that might mean yesterday’s Slack messages, client emails, and meeting notes all get read before the day starts. The agent flags the few decisions actually waiting on you.

For a consultant, the agent might read a call transcript and pull the moments worth a follow-up. It hands you the sharpest pieces, ready to use.

Mine is built for the work I do. At the end of the day, it reads through all my call transcripts. It pulls out the strategic insights and content ideas, tosses them into a workflow, and writes the first draft. When I sit down, the drafts are waiting. I keep the good ones and toss the rest back into pile.

Same machine as the Telegram mess with a completely opposite result. One made noise all day. This one just has the work sitting there when I open the laptop.

Less flashy. Way more useful.

The First Job Is Prep

There’s a real gap between setting an agent up and actually using it. Most people stall at the setup, because the agent never found a real job in their day.

The first agent should make your pile smaller, not bigger.

It pulls the thing that matters to the front. It clears the stale stuff. It bundles the messy stuff so the next step is easy to see.

That’s the job. Make your starting line cleaner.

Your first agent probably lives in the part of your day you already dread.

• The saved links you swore would turn into a post someday

• The call transcript that just get filed instead of mined for insights and content

• The morning scan of messages that get buried as soon as people start waking up

• The weekly client update, where you dig through Slack, email, ClickUp, and your own half-finished notes

When you don’t know why the agent belongs in your day, every setup step feels like extra homework. The moment its job is obvious, the setup is still annoying. But now there’s a prize at the end of it. Even when they hand you a terminal.

The Default Is Actually Impressive…Kinda

The default is to build the most impressive agent you can dream up. (I lovvvvvved the default method)

Memory. Tools. Connections. A dashboard. Maybe a name (mine is named Meatball). Definitely a little personality, because apparently they all have to sound like your favorite celebrity.

The default feels advanced.

It also skips Level 1.

Level 1 is making the first 15 minutes of your day start with clarity.

You open your laptop and the right thing is already waiting. The little agent elves were hard at work while you slept, prepping all of the the store for opening.

Before you ask what agent to build, here are three tests you can run:

The 3 tests for your first agent’s job

• Frequent. It hits most days, not once a quarter. A once-a-quarter task isn’t worth an agent. You want the thing pressing on you over and over.
  Ex: Reading yesterday’s call notes happens every morning.

• Costly. It burns time or focus before your real work starts. This is the tax you pay just to reach the starting line. It’s the warm-up you resent so it probably doesn’t get done.
  Ex: You rebuild last call’s context so it’s ready for the first 10 minutes of the next one.

• Elf-able. The agent can finish it overnight with no choice from you. If the agent has to stop and ask you something halfway, it isn’t done while you sleep. If you are giving it decision-making authority, make sure you do lots of testing first.
  Ex: Running and analyzing sales data from the day before and comparing it week-over-week

If all three are yes, you found your elf’s first job.

It unlocks the door, turns on the lights sets the right thing on the counter.

Then it gets out of the way, so you can do the part that still needs you.

So What’s Your Elf’s First Job?

There is still one teeny problem.

Your highest-leverage prep is almost always something you’ve stopped noticing. You don’t see it anymore. You are too close to your own work (we all are).

The job that would change your day is hiding inside the part of your day you’ve gone numb to.

So you need something that drags it back into view. Either it interviews you until the cost you’ve been swallowing finally has a name, or it reads your actual last two weeks and finds the pattern you’re too close to see.

So I built two prompts that do exactly that. One runs in any chat window. One runs on an agent that can already see your email, Slack, calendar, and transcripts, so its answer is grounded in what really happened, not what you remember. Each one hunts for the single piece of overnight sludge worth handing your elf, and hands you back one job, not a menu.

Subscribe now

The Elf Finder

Two prompts that do one thing: Find the single piece of overnight prep that would change how your whole day goes, and hand you the one job your first agent should open with.

Pick by what your agent can reach today.

If you’ve got a chat window and nothing connected yet, run Version A. It interviews you.

If your agent can already see your email, Slack, calendar, and transcripts, run Version B. It skips the interview, reads your last two weeks, and can build the first artifact on the spot.

Version A: Any chat, no tools connected

Use this when you don’t have a connected agent yet. It works by interviewing you, one question at a time, then picking.

Version B: A connected agent (email, Slack, calendar, transcripts)

Use this when your agent can reach your actual tools. It finds the pattern in what really happened instead of asking you to remember it, and it can build today’s version of the artifact right there. Claude.ai, ChatGPT or Gemini should work, just make sure to toggle on the connectors.

How to audit the answer

The finder worked if it handed you one job, not a list. If it gave you two, it didn’t finish. Send it back: “You handed me two. Pick one, and tell me why it beats the other.”

A good answer sounds like this:

> Your bottleneck isn’t volume, it’s that you rebuild yesterday’s context every morning before you can start. The one job: each night, pull the open follow-ups and waiting decisions out of yesterday’s calls and threads, so a one-screen brief per client is sitting there when you open the laptop.

A weak answer sounds like this:

> You should set up a system to stay organized and on top of your priorities.

Quick note: The video above was our first Live Build / Community Call where I took people behind the scenes in what I’m building and as one subscriber said “a window into your thinking.” Really enjoyed it and will be doing more of them soon.

Last week I taught a four-day Claude Code training. Close to 200 people live each day, 600 registered. Creative strategists and copywriters, most of them brand new to any of this.

Day one wasn’t the fun part. Day one was just getting everybody installed. I wasn’t about to throw a copywriter into a terminal, so we worked inside the Claude desktop app... and even that was a fight. Getting some folks to install the app got a bit dicey.

But somewhere around day two, I noticed something pretty cool.

The site I was teaching from kept getting better, and I hadn’t planned for it to. Every night, it rewrote itself.

Here’s what happened…

After each day, I’d grab the transcript. The whole thing. Plus the chat, plus every question anybody asked. And I’d hand it all to Claude Code and ask:

Based on this, what did people need that I didn’t cover? Is there a better way to structure these resources based on the questions people were asking? What do I need to expand on? What do I need to simplify?

Then I’d tell it to build those resources and drop them into the site.

By the next morning, yesterday’s questions were pages on the site.

Somebody asked about Firecrawl and Exa, two tools I hadn’t planned to touch. So after we wrapped, I said go pull the docs for both and build me a comparison chart, side by side. Next morning, there it was, sitting in the resources.

Another day I recorded a two-minute video after the call, handed it to Claude, and said embed this nicely right here. The people who missed that part live could now watch it sitting exactly where it belonged.

The site got smarter off real feedback. Not off what I guessed people would need ahead of time. And the part that still makes me grin... day three was about building websites in Claude Code. The site I was teaching from was built in Claude Code. So I just pulled up the actual files and showed everybody how the thing they were looking at got made.

Now think about where a course normally lives.

You build it, then you pour it into Kajabi or Circle or Skool, and from that point on you’re working inside their box. Their modules. Their layout. Their idea of what a lesson is supposed to look like.

And sure, maybe you update yours now and then. But you’re doing it within their model. Figure out a better way to teach it halfway through and you’re fighting the template just to rebuild it.

Mine wasn’t capped, and it wasn’t boxed. The site rewrote itself almost in real time. A question from the afternoon’s call was a new page that night, sitting right where it belonged.

I was able to build the exact flow, the exact look, the exact learning experience I wanted. The structure bent to the lesson instead of the lesson bending to the platform.

And before you decide this is some elaborate setup only I could pull off... it isn’t. The whole thing is just files and folders on your computer.

That’s the line I hammered every single day. As crazy as Claude Code looks, strip it back and it’s just files and folders on your computer.

You point it at a folder, and that folder is everything it can see.

You open the same folder in Obsidian for a nicer view... still the same folder.

There are only ever those two moves, and everything happens inside the folder.

So the teaching site was a folder. The transcript went into the folder. Claude read the folder and wrote new files back into it.

There’s no hidden layer.

There’s a folder, and a model that can read and write inside it.

Once you see it that way, the question stops being “could I build this” and turns into “what do I point it at.”

So now you’ve got the idea.

A living folder beats a boxed-in platform, and the engine is a nightly loop. Feed it what actually happened. Ask what’s missing. Let it build while you sleep.

But, knowing that and running it are two different things.

You still need to know:

How to make Claude Code’s output not look like generic AI slop.

How to turn a flat strategy doc into something that looks like your client’s own brand.

How to catch your own mistakes before a client does, without trusting one model to grade its own homework.

That’s what’s below, an upgraded version of what I talked about on the call.

I pulled three resources straight out of that training. Each one’s a thing you can run in the next twenty minutes.

1. The Firecrawl Microsite Recipe
   How I stopped sending Google Docs and started sending interactive sites that look like the client’s own brand, built from their real assets. The exact steps, including the part where it pulled product shots off a retail shelf I never set foot in.

2. The Dual-Model QC Loop
   How I have Claude write a review prompt grounded in the actual codebase, then hand it to Codex to tear the work apart, so two models check each other instead of one model approving itself. Includes the fix for a trap one of my sharper students caught live.

3. The Tools & References Sheet
   The exact sites I reach for... where I get design inspiration so the output doesn’t read as AI-made, where I pull official AI logos, and the one context-window number that tells me when to hand off before the model goes dumb.

So point Claude at a folder. Hand it yesterday. Ask what’s missing.

Then go look at what it built while you slept.

-Max

P.S. I got some really good feedback from the people on the call but after you watch it, let me know what you think.

I am planning to do more of these but like the course, I want to make sure it fits into the box that you want.

Read more

In the summer of 2010, Donna Cooper took her family on a day trip to Death Valley. It was a good day. Long, hot, the kind you end tired and know you’ll sleep good. When it was time to head home to Pahrump, she did what we all do…she asked the GPS for the shortest way back.

Donna called the GPS Nell.

Nell said go 550 feet, then turn right. At 550 feet there was a little path. Then a quarter of a mile and turn left. There was nothing there. Just desert, in every direction, in one of the hottest places on the planet.

“She had me running in circles for hours and hours and hours,” Cooper said later.

Three freakin days.

That is how long the Cooper family was lost. A search and rescue helicopter finally spotted them and luckily everyone survived.

The rangers in Death Valley call this “death by GPS” and it actually happens a few times a year. Around the same stretch, a mother and her son followed their GPS down an abandoned mining road and got stuck for five days. She barely lived but this time, the boy did not.

What unsettles the rangers is that the device is almost never wrong. It is too right. It computes the most direct line from where you are to where you want to be, and it sends you straight down a road that stopped existing forty years ago, with total confidence, in a calm voice, the whole way.

Here is the crazy part…

After three days in the desert, Donna Cooper did not lose faith in GPS.

She bought a new one. She named it Rosie.

We talk about trust like it is a decision. Cooper decided to trust Nell. Then she decided to trust Rosie. But that is not what happened, and you know it is not, because you have sat in that passenger seat too.

The trust was not a choice she made. It was a feeling the device created in her.

A confident voice, a clean line on a screen, an arrival time down to the minute.

And the feeling of being guided is so much stronger than the evidence of the road dissolving in front of you that the feeling wins, every time.

The tools that put you in danger are not the ones that fail you. A tool that fails you, you stop trusting. The dangerous ones are the ones that make you feel guided while they walk you somewhere you would never have driven with your own eyes open.

The Guided Feeling

You have had the guided feeling. The draft that appears in eight seconds, so smooth it reads like it was always there. The little clock in your head that says this used to take an hour, and now it takes ten minutes, look at me go!

And the feeling is real.

But what we are all forgetting to ask is whether the feeling is attached to anything.

And the problem is not the skeptics who refuse to touch the stuff. They are fine. They kept their eyes on the road.

The problem is the rest of us.

The experienced ones. The people who use AI well, feel sharp doing it, and have no idea the road stopped existing two turns back. We are producing more than we ever have.

But some of us are getting slower while we do it, and have no idea at all.

Subscribe now

The Clock Disagrees With You

In early 2025, a research group called METR ran the study you would run if you were being honest with yourself. It’s like when you haven’t gone to the dentist in years and keep putting it off because ya just would rather not know.

They took 16 experienced developers. These were people working in their own codebases, projects they had contributed to for an average of five years, some with more than a million lines of code. They gave them 246 real tasks from their actual work.

Then they flipped a coin on each task:

AI allowed, or AI not allowed.

Before they started, the developers predicted AI would make them about 24 percent faster.

After they finished, they reported that it had made them about 20 percent faster.

The screen recordings said they were 19 percent slower.

No, not slower than they hoped. Slower than they were without the tool at all. The thing they reached for to speed up was the thing slowing them down, and the real kicker, the part that should really make ya feel funny, is that even after living through it, they still walked away believing the opposite.

The Confidence Is Wired Backwards

It goes one level deeper.

Microsoft and Carnegie Mellon surveyed 319 knowledge workers about when they actually think critically while using AI. The finding ain’t good, but honestly, makes a ton of sense.

The more you trusted the tool, the less you thought. High confidence in the AI predicted less scrutiny of what it gave you. High confidence in yourself predicted more.

Read that again, because it is doing something quiet and nasty. The better the tool feels, the more it sands down the exact reflex you would need to catch it being wrong.

I heard a story of a founder who knows his numbers like the back of his hand. He ran a board document through AI and forwarded it. A partner caught a figure that was off by a year. He had read it.

He just had not read it in the way he reads everything that does not arrive pre-formatted and “certain”.

The polish bought his trust before his judgment ever showed up to work.

You get in the habit of feeling like it is handled but it costs you the habit of checking, at the very moment checking is the whole job.

I catch myself in this moment all the time. These tools have gotten good. Real good. I will be working late at night and realize I have 3 Codex sessions, 2 Claude Code and a few Chrome browsers humming along.

It’s so much easier to keep tapping “approve” and trusting the AI and man…it feels damn good making that much progress.

But as one of my favorite sayings goes, “the bill always comes in the end.”

The Muscle Is Already Going

This is the one you cannot feel at all, because it is measured underneath feeling.

MIT’s Media Lab wired up 54 people with EEG and had them write essays.

One group used ChatGPT.
One used a search engine.
One used only their own brain (crazy I know).

They watched the brain activity the entire time.

The brain-only group lit up the widest with the most connected networks.
The search group was less.
The ChatGPT group showed the weakest neural connectivity of the three.

The cognitive engagement scaled down in lockstep with how much the tool was doing.

It’s the cognitive version of taking the escalator every day and then being surprised, months later, that the stairs leave you winded.

But here is the most surprising/unsurprising part…

The AI group could not quote their own essays. Sentences they had produced minutes earlier…couldn’t recall them. Because in a real sense they had not written them. They had watched them get written.

The researchers call this cognitive debt.

It feels like nothing and that’s precisely the problem. The bill arrives later (see!), in the one moment you need to think without the tool, and reach for the muscle that’s no longer there.

Why Your Gauge Stopped Working

For your entire career until about two years ago, the feeling of being good and the fact of being good rolled together.

You felt fast because you were fast.
You felt sharp because you had done the reps that morning, that year, that decade.

The feeling was a trustworthy gauge for one boring reason.

There was no way to feel productive without actually being productive. Producing the work was the proof. They were the same event.

But AI has started pulling them apart.

For the first time you can have the complete sensation of expert output with almost none of the underlying work, and your internal gauge has no idea anything changed. It still thinks it is counting reps…but it’s counting vibes. The needle that ran your whole career came loose from the engine, and it still swings, confident as ever, pointing at nothing.

Your Company (Or Your Client) Is Running the Same Thing

Now zoom out to a team, or to a client you advise.

A group ships 30 percent more this quarter. Leadership reads that as a win and raises the number for next quarter. Nobody measures the thing that does not show up in the dashboard:

Whether the people producing the work still understand it well enough to defend it, fix it, or rebuild it when the tool is down.

The organization optimizes the one thing it can see and goes blind to the one thing it cannot.
The startup asks, what can we produce now. The question that would actually protect them is the one nobody asks out loud because it’s not convenient now.

What can we still do when it stops working? Because we all know it will stop working eventually.

A Test Worth Running This Week

So here is something to try.

Pick a task you always hand to AI. The one you have fully outsourced because at this point it would feel stupid not to. Do it once this week without the tool. Time both versions.

Then notice two things.

The clock, which tells you whether you were actually faster.

And the harder one. The parts you could not do as cleanly as you remember being able to.

That second reading is important. The stopwatch measures speed.

The struggle measures the muscle.

You are not doing this to feel bad, or to go back to grinding everything out by hand like a monk. You are doing it to reconnect the gauge. Use AI to move faster through the parts that do not require you. Stay inside the parts that do. Keep your eyes on the road even while ol’ Nellie is talking.

Watch It Run Backwards

Let me show you the reverse happening.

I know a Fractional CMO. Let’s call him Reza. He runs marketing for three companies at once and is sharp, fast, and exactly the kind of person who should be using AI hard. And boy was he. Positioning memos, campaign briefs, board updates, all of it drafted in minutes by asking Claude to write the thing and then cleaning it up.

Then a founder he works with stopped him in a meeting. Pushed on one line of a positioning memo. Asked him why he had recommended going after the mid-market instead of staying upmarket. A simple question. The reasoning should have been right there.

It was not there.

Reza knew the memo said it. He could not reconstruct why it said it, because in a real sense he had not decided it. He had watched it get decided. He recovered, the way we all do, but he drove home knowing exactly what had happened. The gauge had been reading full the whole time the tank was draining.

So he changed one thing. Not the tool. The order.

Now, before Claude writes a word, he opens the first prompt in the pack and it refuses to draft anything. Instead it makes him talk first. He brain-dumps his actual take, messy, into voice mode for ninety seconds. Then the second prompt does not write the memo either. It interrogates him.

“You said mid-market. What changed your mind from the upmarket play you ran last quarter? Name the specific signal. If you cannot name it, say so and we will flag it as an assumption, not a finding.”

Reza answers out loud. The thing that used to be invisible, the actual decision, gets dragged into the light where he has to stand behind it. Only then does the third prompt sharpen what he built, and it flags two lines where his reasoning is still thin and tells him so instead of papering over them.

Same memo with roughly the same words…Maybe ten minutes slower.

“But now it’s mine,” he said. “I could feel the difference in the next meeting. When she pushed, I was already standing there.”

That’s it. Not doing it the hard way. Doing it in the order that keeps you grounded while it happens.

What the Test Can’t Do

You now know the gauge is disconnected. You know which group you are in, the one that drafts and stays inside the thinking, or the one that generates and copies. And you have the one test to feel it for yourself this week.

Here is what the test cannot do.

It cannot rewire your default. The moment a real deadline lands and the tool is sitting right there, helpful and instant, willpower loses. You will reach for the fast version, because feeling fast is the one thing that never feels like a cost. Reconnecting the gauge in your live daily work is not a decision you make once. It is an order of operations you install so the thinking happens whether or not you are disciplined that day.

Your Competence Is Reading This Right Now

And it is already filing objections.

It will feel like:

That study was just developers, my work is different.

It will feel like:

I am the careful one, I always check the output.

And the last one, the quiet one:

I can tell the difference between real speed and the feeling of speed.

That last one is the exact belief those developers were holding while the stopwatch called them liars.

Feeling like you can tell the difference isn’t the same as being able to. And that confidence is the first thing AI takes from you. Not by making you feel worse, but actually by making you feel better.

Every serious practice ever built knows this and builds around it. Sprinters do not trust how fast the race felt. They trust the tape. Musicians record themselves because the playing always feels better than it sounds. The move at the center of mastery, in every field, is the same.

An external instrument, brought in on purpose, because the internal one lies in your favor.

You’ve lived with the gap between feeling good and being good your whole career. The difference now is that something finally learned to sit inside it, speak in a calm voice, and give you directions you’ll actually follow.

Donna Cooper named hers Rosie.

What’s yours called?

-Max

Knowing the gauge is loose won’t reconnect it.

That is what The Sanity Pack is.

Four prompts you paste into Claude or ChatGPT that flip your default from generate-then-copy to think-then-sharpen, the move that kept the MIT study’s second group with their brains on the line.

You get the four prompts, a voice-mode version for thinking out loud the way Reza does, a one-page weekly self-audit to catch yourself drifting, and the full worked example.

Run them on one real task this week. If they don’t catch something you would have shipped and couldn’t defend, reply and I’ll refund you. Keep the prompts anyway.

It’s below.

Read more

Not “a little rusty” and didn’t “quit piano as a kid.” He has said, plainly, that he has no technical ability in music and knows less about the mechanics of recording than almost anyone who has ever done his job.

For long stretches of a session he looks like he’s snoozin’ on a couch with his eyes closed.

But this is the man who co-founded Def Jam out of his NYU dorm room and then spent four decades producing records that don’t feel like they should even be on the same shelf.

• Run-DMC and Slayer

• Red Hot Chili Peppers and Adele

• The Beastie Boys and Johnny Cash

Genres that basically hate each other, all routed through the same guy on the same couch.

Musicians who work with him for the first time tend to ask a version of the same question…

What, exactly, do you do here?

The honest answer doesn’t make them feel any better. He listens, and he tells them the truth about whether it’s any good.

That’s the whole job. He cannot make the sound. He can only tell you, with a confidence that has been right for forty years, when the sound is true and when it’s filler pretending to be a song. The artists can all play. Some of them are the best in the world at it but what they hired was that magic ear.

We treat that as an exception. The genius who “can’t work the gear.”

But there have always been two different roles hiding inside the word “skill,” and we spent about a century thinking they were one.

One job is operating the tool. Fingers on the keys, hands on the board, the craft of making the thing come out.

The second job is knowing what the thing should be. Whether it’s any good. Whether it moves. Whether to keep it or kill it.

Almost everywhere, we bundled those two jobs into one person and called the bundle “skill.” You earned the right to judge the work by first proving you could do the work. The ear came after the hands.

But our guy Rubin skipped the hands. He went straight to the ear. And it turned out the ear was the rare part all along.

Rubin will tell you music is almost incidental to what he actually does. He happens to work in it. Knowing what’s good is the same in any room.

I know, I know. This is kind of an AI newsletter so you can probably guess where this is going.

For most of your life, one sentence kept you out of an entire category of work.

“I can’t code.”

One, maybe two years ago, that sentence meant software was a thing other people made, and you described what you wanted to them, badly, in a doc that cost you thousands.

That barrier has been moved by millions of tiny robots, pre-trained on the best coding practices on the internet.

The operating layer, the syntax, the part where you translate an intention into something a machine will actually run, is the part the robots do now.

You say what you want in plain English and it builds. Not perfectly and not always right on the first try. But it builds, and it never once needs you to know what JSON is.

What it cannot do is the second job.

It cannot tell when its own work has no pulse. It hands you something that it is VERY proud of and…average. And then it waits. It waits for someone to look at the output and say “this is fine!” and fine is the problem.

Rubin puts it better, with guitars. A lot of records get built by stacking the same part over itself, double it, triple it, until you’ve got a “wall of guitars”. And when there’s a wall of guitars, you hear guitar. You don’t hear a person playing guitar. It goes generic.

The personality in the one take where you can hear the fingers catch the strings.

The machine(AI) hands you the wall of guitars. Every time. Full, competent, and nobody in the room.

That someone is you. The job that’s open is the ear.

And the people least likely to notice the job is open are the ones who spent ten years building the exact ear it needs.

Watch how “code is the skill” quietly stopped being true.

The first thing you notice is the swap. You used to need a translator, someone who could take “make it feel premium but not stiff” and turn it into the thing that runs. Now you type “make it feel premium but not stiff” and the translator is built in.

This feels like the technical person still has the edge. They speak the machine’s language natively and you’re working from a phrasebook. So you defer. You keep treating the person who can operate the tool as more valuable than the person who knows what good looks like. That deference is the first cost, and it’s already out of date.

Go one level down. The hard part of building anything was never the typing. I learned this the slow way. The hard part was deciding, precisely enough that the result could be judged, what you actually wanted.

I watched a strategist with no technical background pull better work out of Claude than a developer sitting two seats away. The difference was simple. She could say exactly what “good” meant for an ad in her category, and he couldn’t. He could build whatever she specified. He just had no specification of his own.

When that happens to you, you’ll blame the prompt. You’ll go rewrite it a sixth time. The prompt was never the thing in the way. You hadn’t decided what good was, and no amount of rephrasing decides that for you.

Claude Code Accelerator

None of this is new. Rubin proved it at eighteen. When he started, the only hip-hop records out were made by studio pros who built other kinds of music for a living. They had all the craft. They didn’t know the club. So they made a clean, polished version of a thing they didn’t understand, and it came out fake.

Rubin could barely record. But he’d stood in that club every week and knew exactly what it was supposed to feel like, so he made the real one. The craft was all on the wrong side of the glass. The scarce thing was knowing what good was.

Now the bottom level, the one that's harder to look at directly. Specifying what to build was the floor above. This is the floor beneath it.

Whether what came back is any good, and the nerve to say so when it isn't. When the machine can build anything you describe, that judgment is the only scarce thing left in the room.

Rubin has a word for his actual work.

Reduction.

He coined it at nineteen, on an early LL Cool J record, where he printed “Reduced by Rick Rubin” instead of “Produced by.” Production meant building up. What he was doing was the opposite. Stripping a thing down until only what matters is left, and knowing, in your body, where that line sits. That knowing is the real work. It is the instruction set now. Your taste is the thing that tells the AI what to make, and then tells you whether it made it.

This feels like you’re “just giving feedback.” Like the build is the work and your notes are commentary on the work. It goes the other way now.

The notes are the work. The build is the commodity.

For a hundred years, taste sat on top of craft like a penthouse. You took the elevator up through technique to reach it. That’s why it felt like a luxury, the finishing layer you added once the real skills were in place.

The elevator is free now. Anyone can be at the top floor by the afternoon.

So the penthouse isn’t a luxury anymore. It’s the whole building. We are watching the floor and the ceiling swap places in real time, and the people most exposed are not the ones who never learned to code. They’re the ones whose only edge was that they could.

You can see the same confusion at the company level. Every job post wants someone “proficient with AI tools.” They’re screening for operators, people who can drive the machine.

The thing they actually need is harder to test for and worth far more. Someone who can look at what the machine produced and know, fast and correctly, that it’s wrong.

What matters is whether you can tell when it’s lying to you. Using it is the easy part.

Something to try this week. It takes about ten seconds and it works on any AI output you touch.

The next time AI hands you something and you feel the little reflex to fix it, stop your hands. Don’t fix it yet. Ask yourself one thing.

Could I write down the rule I’m using right now?

You felt the thing was off. That feeling is a rule you already hold, sitting just under language.

“The opening is too eager.”

“It’s explaining the joke.”

“Real people don’t talk in lists.”

When you reach in and pull that rule up into a sentence, you’ve made a piece of your taste visible that was invisible a second ago.

Do that for a week, every time you correct an output, and you’ll notice something uncomfortable and good. You have hundreds of these rules. You’ve had them for years. They’ve just been trapped in your head, leaking out one correction at a time, teaching no one and saving you nothing.

Your taste is reading this right now. It’s the thing that has been deciding, paragraph by paragraph, whether this was worth your time.

You’ve been treating that faculty as the cheap part of your work. The part you’d happily hand off so you could get to the “real” output.

It will feel like the soft skill.
It will feel like the thing you can’t put on an invoice.
It will feel like the part that doesn’t count, because you didn’t sweat for it the way you sweat for the technical stuff.

Rubin lay on that couch for forty years being the most valuable person in rooms full of virtuosos, because he understood the thing they kept forgetting. Anyone in that room could play. He was the only one who could hear.

The machine can play now. All of it. Every instrument, every genre, on demand, for nearly nothing.

The hearing is still yours.

It might be the only part that was ever really yours.

-Max

P.S. Putting that hearing somewhere it can actually work, instead of leaving it trapped in your head leaking out one note at a time, is the thing I built an entire workshop around.

Claude Code Accelerator starts Wednesday. I built it with Luke and Mario, two of the sharpest copywriters and creative strategists I know. Which is to say, two people who already do the thing this whole piece is about for a living.

If you already know what good looks like and you’ve been waiting for permission to build, this is it.

Claude Code Accelerator

Between 1883 and 1929, Andrew Carnegie paid for 1,689 public libraries across the United States.

He’d grown up poor in Pittsburgh. As a teenager working as a messenger boy, he borrowed books from a private library a local colonel opened to working boys on Saturdays. Carnegie never forgot it. When he sold his steel company and became the richest man in the world, he decided to give towns the thing that had changed his life.

Then he attached a strange condition.

Carnegie would pay for the building. The bricks, the shelves, the reading room with the tall windows…

But he wouldn’t pay to run it.

Before he released a dollar, the town had to pass a resolution, a public and binding promise to fund the library every year afterward, usually at ten percent of what the building had cost. The staff, the heat, the new books. Forever.

Towns pushed back. They wanted the gift without the obligation. A library was a fine thing to have on the main street. But Carnegie had seen what happened to libraries a town owned and wouldn’t run. The doors stayed locked five days a week. The shelves thinned. The building turned into a monument to a thing the town didn’t actually use.

So he refused to fund those at all.

Carnegie had worked something out. The check buys a building. Only the town can buy a library. So he made them buy it first, in writing, before he spent a cent.

The Bike in the Corner

Getting a thing and running it…really committed to it…are different jobs.

Your friend buys a Peloton in January. The truck shows up Saturday. Two guys carry it up the stairs, plug it in, set up the screen. The Peloton has officially arrived. By June the bike is in the corner of the bedroom holding a sweatshirt and a phone charger. Your friend still owns a Peloton. Your friend is not a person who rides a Peloton.

Carnegie knew this. He’d seen the locked library. He understood that putting a thing within reach of people and getting those people to actually run it are two different projects, and the second one is the hard one.

The first is a check.

The second is a commitment, and nobody can write it on your behalf.

Subscribe now

Everyone Got One

A few days ago, Google ran the largest distribution of AI agents in history.

At it’s I/O conference it announced that the agent is now the product.

Gemini rebuilt into the search box. Gemini Spark, an assistant that “Give it a task and it works in the background 24/7, even if your phone and laptop are turned off. It operates autonomously, but always under your direction. You choose to turn it on and it's designed to check with you before taking major actions.”. Agentic features reaching the 900 million people who open the Gemini app and the billion more who meet AI inside Search. The keynote framed it plainly.

You have an agent now. Everyone does.

Google built the building. It’s a very good building. I built a smaller one this week. A recap of every I/O announcement, a Omni prompting skill and a few other goodies.

But here’s the question Carnegie would ask before he believed any of the hype…

Who’s going to run it?

I am not talking about the skeptics. The skeptics were never the problem.

I am talking about what happens when everyone gets behind this agentic-wheel without really understanding what it can do?

This is no longer just an LLM chat. It can take actions “on your behalf.” But buyer beware because we are about to have a library with no librarian, and can’t quite tell.

Three Reasons You Can’t Tell

Having the building feels like running the library for three reasons, and each one is harder to see than the last.

1. The agent arrives friendly

   It greets you by name. It offers example prompts. You try one and it does something impressive on the first attempt.

   Feels like: It’s working.

   The cost: Those example prompts came from Google

   None of them came from you. Look at what people are actually sharing online two days after the launch. The clever Gemini Omni trick going around, the one that turns a butterfly into a swarm of fireflies, is Google’s own example, lifted straight from Google’s own prompt guide.

   
   The crowd is copying the manual and posting the manual back at each other. You’re running the company’s demo and calling it your workflow.
   

2. The agent is generic by default

   
   It knows everything about the world and nothing about you. It can write an email. It can’t write the email you’d send to the client who went quiet last month, in the register you use with that account, about the thing only you know happened on the last call.

   Feels like: It’s capable

   The cost: It can do anyone’s work. It can’t quite do yours

   I watched this play out at a marketing agency. The founder couldn’t scope a new project without sitting in the room. Every statement of work and proposal needed his judgment. His team had the exact tools he had, the same subscription, the same models. The tools still couldn’t scope an SOW or a proposal, because scoping those were never generic tasks.

   It was his specific patterns, built over years, living nowhere but his head. I pulled that pattern out of 7 of his past SOWs and proposals and built it into a system. Six hours of his time per project became 30 minutes. The team does this without him now. Nothing about their AI access changed. One task got installed and that was the whole difference.

3. You shrink to fit the tool

   
   This is the one you can’t see without help. You adapt to the generic version. You learn which questions it answers well, and you quietly stop asking the ones it doesn’t.

   A tourist with a translation book can order coffee and find the station. She’s got a few hundred handed-down sentences. She’ll never have the conversation that actually mattered, and after a while she stops reaching for it, because it has trained her to only ask what it can answer.

   Feels like: I’ve adopted AI

   The cost: You adopted the demo. The work that needs you, that runs on your judgment and your context, never gets touched. And you stop noticing it’s missing

The three reasons have been here all along.

The Damage Used to Stay Inside

This used to be a smaller problem.

When AI was a chatbot in a separate tab, running the generic version cost you some speed and some polish. You got a mediocre draft. You fixed it. The damage stayed inside the document.

The agent changes that. Agents act. They send the email. They book the slot. They move the money. They make the call you’d have made, at whatever quality you set them to.

The mediocre version goes out under your name, gets read, gets remembered. By the time you notice the tone was off for that client, the position has already been taken in their head. You can apologize. They can’t unread. Distribution put that in 900 million hands this week. Installation (using the thing) is the only thing left standing between capability and consequence.

The Question Nobody’s Asking

I am the last thing from an AI doomer. If you follow me or have read any of my content you will know I love AI, I think the Gemini and Google updates are freakin awesome.

But the tech is moving SO much fast than most people’s “ai-literacy”. They are being handed a powerful set of tools and instructions, incentives and often work mandates to use it as fast as possible.

This sure doesn’t feel like it will end well…

Picture it inside a business.

A company that rolls the generic version out to 50 people has done something that looks decisive and isn’t. A license on every seat. A leadership team asks whether everyone has access and a training session forgotten in a week.

But here is the sneaky part.

It’s going to sound good. It’s going to look good. It’s going to create PowerPoints for you in minutes that used to take you hours and your leadership is practically begging you to do it.

Now, if that AI hasn’t been loaded and trained on you, your company, your situation, your thinking and your judgement it will give you the same answers it gives everyone else.

Answers that, on the surface, will make you think you will get that next promotion in no time.

Until you try to explain them.

When you have to get up in front of the team and present that fancy pants deck with all those animated charts and realize they are completely shallow.

28 slides that are directionally right but operationally useless (one of my favorites terms). Without the proper context and without people moving from knowing to understanding AI, this will happen over and over again.

So here’s the question that hasn’t been answered.

Sit with it. Write the answer down. Be strict.

If the honest answer is a faster email, a cleaner summary, a decent first draft, then the agent did for you the exact thing it did for 900 million strangers.

If the answer is specific, a draft only your agent could write because it knew your client, a risk it flagged because it had learned your blind spots, a call it got right because it caught what you would have caught, then something is installed. You’ll know, because a stranger couldn’t have gotten it.

Ask it this Friday. Then every Friday for a month. The weeks the answer comes back thin will teach you more than the weeks it’s good.

Twenty Years Across the Table

Financial advisor, 20 years in. Let’s call him Glenn. He had ChatGPT, he had Claude, he had every tool the headlines told him to own. He used them the way the tourist uses the translation book. Summarize this. Clean up that. Distribution, working exactly as distributed.

Then we installed one thing.

Glenn’s edge was never the financial knowledge. Every advisor has that. Glenn’s edge was reading a room. Twenty years across the table from couples had taught him to hear the thing under the thing, the hesitation a client would never say out loud. That skill lived only in Glenn. It had never been written down. It was the reason his clients trusted him, and the reason he couldn’t be in two meetings at once.

We extracted it. The patterns, the tells, the questions he asked himself in the car after a hard meeting. We built it into a system that reviews his call transcripts the way Glenn’s own instinct would.

He uploaded a transcript from a meeting he thought had gone fine.

The system flagged five signals he’d missed. A couple who’d said yes with their words and no with everything else. It told him the close he was planning, a gentle bit of pressure, would walk them straight out the door. It told him to do the opposite. Slow down. Name the hesitation out loud. Ask the question he’d have asked himself if he hadn’t been tired.

Glenn made one call. He followed the deeper read. The couple rescheduled within 20 minutes.

A generic agent could have summarized that transcript for anyone. It couldn’t have caught what Glenn would have caught, because catching it required being Glenn. For one afternoon, it was.

Why You Can’t Just Do This

That recap I created and linked above took an afternoon. It’s free because it should be free. A recap tells you what shipped and changes nothing about whether you can run it. The running was always the scarce part. The running is the only thing worth paying anyone for.

Carnegie wrote his clause into the contract because he knew the building would lie to the town. A library looks finished the day the roof goes on. It isn’t finished. It isn’t even started. It starts the first morning someone unlocks the door, and it has to start again every morning after that.

You understand the difference now.

• Distribution is the building.

• Installation is the librarian.

You can feel which one you’ve been running.

What you can’t do yet is the install itself. It’s genuinely hard, and not in the way you’d guess. Nothing about it is technical. The hard part is that installing your own expertise asks you to be your own interviewer, to pull a pattern out of your head while your head is busy being the expert. You can’t easily extract the thing you’re using to do the extracting.

That’s what The Personal Pattern Protocol is for.

It’s a three-part prompt sequence you run inside Claude, ChatGPT, or Gemini. It does what we did with Glenn, on your own expertise, with no consultant in the room. Glenn’s installation, above, is the worked example. The three prompts and the weekly audit are below.

The Personal Pattern Protocol

Three prompts. The first two are the setup. The third is what keeps the install alive.

The setup takes an afternoon. The audit takes ten minutes, every Friday, forever. Starting this one. The protocol has the same shape as the rest of the article. The setup builds the library. The Friday audit is the running.

Before You Start

Pick your tool. Use whichever AI you already pay for. The interview works the same in Claude, ChatGPT, and Gemini.

Pick your task. Three tests have to pass. Other people rely on you for it. You can describe a recent real instance of it from start to finish. Someone replacing you would do it noticeably worse. If several tasks qualify, pick the one where the “noticeably worse” is most obvious.

Prompt 1: Extract

This prompt interviews you. It will not help you with your work. Its only job is to pull one piece of your judgment out of your head and onto the page.

Read more

You wait. A minute. Two. Five.

There’s no way he’s still down there!

This is the Bajau, Southeast Asian sea nomads who have spent centuries living on boats, visiting land only for supplies, and spending five hours of every working day underwater. The most extreme of them can dive 200 feet on a single breath and stay submerged for thirteen minutes.

And for a long time, nobody could explain how.

In 2018, a Berkeley PhD student named Melissa Ilardo arrived with an ultrasound machine and started doing scans.

The Bajau spleens were 50% larger than those of their non-diving neighbors. She sequenced the DNA and found a specific mutation in the PDE10A gene. The Bajau had evolved a biological scuba tank inside their bodies, an oversized organ that stored extra oxygenated red blood cells and released them during dives.

They’d been carrying it inside them for generations and they had ZERO idea.

You too have a superpower that has been invisible to you your whole life. But you can’t see it because you’re too close to it.

I call this your Cognitive Fingerprint™. The signature patterns underneath your expert work. These are often invisible to you, but your transcripts can bring them to the surface…if someone knows what to look for.

That is exactly what happened with a financial advisor I’d spent weeks building a system for. He had no idea what was inside him or his transcripts, either.

Subscribe now

He Started Typing Something I Never Taught Him

We fed twenty-some client meetings through the analysis. Pulled patterns out. Configured prompts to his specific sequences, decision DNA and even his blindspots.

One morning, he was on a Zoom call with me, poking around and testing things.

Then he stopped clicking and started typing:

“This prospect just ghosted me on our 2nd presentation after I thought we nailed it. Before I send my usual follow-up, can you look at this transcript and see if there were any red flags I was missing?”

That wasn’t in my documentation. That wasn’t part of any setup walkthrough. He just... felt it. I sat there watching him troubleshoot his own sales process with an AI that actually understood how he sells. What happened next was pretty darn incredible.

Eight Times Per Meeting

His system knows things about him that he didn’t know about himself.

• It knows he says “How are you for time? Any questions before we move on?” eight times per meeting. Eight. I showed him the analysis and he laughed out loud. He’d never counted.

• It knows his Bucket Framework. Soon, now, later. And that he runs the same sequence every single time without variation. Time horizon first. Then purpose. Then products. He thought he was improvising. He wasn’t.

• It knows the spending permission reframe. This is the thing that makes him different from every other financial advisor running the same financial playbook.

And most importantly, when it has the full picture, it knows “why” he does all of these things and can give him “obvious” recommendations that he couldn’t see himself.

In the scenario above, AI told my client there were multiple red flags during the first meeting that indicated the couple was nervous and “do not send” his normal follow up email. They probably already feel guilty about not showing up and calling them out would just send them into the hiding.

Based on the recommendation, he pivoted and sent a completely different message than we was planning to.

“They called back an apologized 20 minutes later and were back in the office one week later! Incredible!”

This goes beyond just a one-time, “I need to save this account” situation.

For instance…

When projections show a client leaving $5 million to their kids, most advisors nod approvingly. He catches them by surprise and says “this means you could spend more now.”

His clients aren’t just afraid of running out of money. They’re afraid of dying before they enjoy retirement. Their spouses keep asking them to slow down. Take a trip. Actually live a little before it’s too late.

And then their financial advisor, the person they expect to lecture them about saving more, tells them to spend.

You can watch the tension leave their shoulders. According to him, “I don’t know, that is just what I do.”

But here’s what those patterns look like when you pull them out and document them:

• The “Flexibility Emphasis Pattern” on the left. Five steps he runs every time a client freezes up about making a permanent decision.

• The “Education First Pattern” on the right. His sequence for teaching before asking.

He didn’t write these. I extracted them from his transcripts. He runs them automatically, the same way every time, and had no idea how consistent he was until someone showed him. When he needs to teach his junior advisors his process, the level of sophistication, customization and depth goes way beyond anything he’s done before or was able to generate from just prompting ChatGPT.

The 90% That Never Made It onto Paper

This is why expertise can’t be extracted from what people say about themselves. Books, courses, interviews, biographies. They all capture the explicit stuff. The 10% that made it onto paper.

The other 90% never got documented. It’s sitting in the nervous systems of people who’ve been doing the work for decades.

There is nothing new about the phenomenon. Cognitive scientists have been mapping it for sixty+ years.

• Michael Polanyi called it tacit knowledge

• The Dreyfus brothers tracked how it develops

• Gary Klein watched experts make impossible decisions on instinct

Their conclusion came down to the same thing.

When things look familiar, experts don't deliberate. They execute.

Tacit knowledge (the execution) is the methodology hiding inside your transcripts. The patterns you run without realizing you run them.

But here’s the thing about this elusive little concept…

You can read everything ever written about tacit knowledge and still not have any.

You can’t prompt your way to it.
You can’t even ask the expert to explain it.

Polanyi himself said,“We know more than we can tell.”

You have to earn tacit knowledge yourself. There’s no shortcut.

And even after you’ve earned it, that sucker still refuses to let you know it’s there.

545%

A company wanted to figure out why some of their salespeople closed deals at 2-3x the rate of everyone else. They were using the same product, scripts and in the same market as their colleagues. But they were getting wildly different results.

So they studied them. Thousands of’em.

Almost half the managers and observers had no idea what made their best people the best.

So Cresta pointed AI at the transcripts.

• Not asking anyone what they did.

• Watching what they actually did.

They found four specific micro-behaviors that separated the top performers:

1. Setting expectations

2. Asking discovery questions

3. Handling objections

4. Closing

But the real hero was discovery questions. On average, top performers asked 48% more and on certain types )needs-focused, upsell) the gap widened to 545%.

Five hundred and forty-five percent.

These patterns were invisible to everyone, including the people doing them. The behaviors were so automatic that nobody could see them until AI surfaced them from the transcripts.

The Four Boxes He Checks When a Deal Dies

The financial advisor has his own checklist now.

At the end of each day his transcripts are auto-processed through his AI system, loaded with his Cognitive Fingerprint scanning for his superpowers in action. It’s helping him:

• Scale his expertise without slowing down: Transforming his daily, unconscious habits into a tangible, growing library of training assets without requiring him to spend hours lecturing his team.

• Decode his own “invisible” moves: Pinpointing the exact moment he successfully deployed a technique like “The Value Reframe” or “The Permission Check,” isolating the precise phrasing and timing that made the client say yes.

• Feed his “training engine in a box”: Automatically generating fresh, one-page Advisor Training Cards based on real-world scenarios from that exact day, rather than relying on stale, hypothetical theory.

• Standardize firm-wide excellence: Giving less experienced advisors a daily, updated playbook of recognition triggers, success indicators, and exact quotes they can review before their own meetings tomorrow.

• Audit his own consistency: Providing a daily mirror to ensure he is continually executing his high-scoring patterns—like “The Education First” and “The Spending Shift”—at the highest possible level.

• Protect and codify his intellectual property: Systematically capturing the unique, proven frameworks that drive his firm’s revenue so they become permanent, transferable assets rather than just ideas stuck in his head.

That’s the system surfacing what was inside him from the work he was already doing.

What Would Your AI Tell You?

What would your AI tell you about your last five calls?

Probably something useful. The kind of feedback most AI can give about most conversations in your industry.

Now imagine asking a different question.

“This one didn’t convert. Check it against my methodology. What did I miss?”

That requires something different. A system that knows your patterns. Your sequences. The micro-behaviors you’ve been running for years without realizing how consistent you are.

One reads the transcript. The other reads you.

You can see what he built. What you can’t see is how to build it for yourself.

THE THIRTEEN-MINUTE METHOD

Three prompts that turn your transcripts into a closing edge you can run against any conversation, forever.

The Bajau go down with intent and come back with what they went for. Same idea applied to your transcripts.

How to Use This Kit

Read more

In the late 1990s, if you wanted a job at Microsoft, you sat across from someone who asked you how you would figure out how much Mount Fuji weighed. The riddles weren’t really about Mount Fuji.

They were a proxy for the skill Microsoft thought it was hiring for at the time, which was the ability to reason through a problem you’d never seen before with no resources except your own head.

Google’s interviews famously took the same idea further. Candidates wrote algorithms on a whiteboard while three engineers watched them reason out loud. McKinsey ran case studies. Investment banks ran modeling tests.

Every one of those interview formats was a bet about what the company actually needed from a new hire... encoded as a test you could either pass or fail in a small room with the door closed.

Tests changed when the work changed.

The whiteboard test stuck around for thirty years because writing code by hand while reasoning out loud was a reasonable proxy for the actual job: thinking on your feet while producing. The case study stuck around at McKinsey because the actual job was structured ambiguity. When the work shifted, the tests shifted with it. Sometimes faster than the people taking them realized.

I came across a Sam Altman quote that lands differently once you understand what those tests were really filtering for.

“We basically would like to sit you down with something that would have been impossible for one person to do in two weeks this time last year, and watch them do it in 10 minutes or 20 minutes.”

To compress two weeks of work into twenty minutes, you have to be working alongside AI. The bottleneck shifts to how clearly you can describe what to produce.

OpenAI is changing what they’re testing for. The framing in the press was that they’re “evolving” the process, which is a polite word for what just happened. They’re not testing if you can build something anymore. They’re testing if you can describe what to build clearly enough that AI does it for you. The build follows. The description is what they’re grading.

The company making the most advanced models on earth doesn’t care if you can write the code. They care if you can specify the work.

Subscribe now

OpenAI isn’t the point

I know what you’re thinking.

“Max, why should I care? I have zero interest in working for OpenAI.”

I get it. But you’re missing the point.

OpenAI made the shift explicit. If you produce real work with AI, the same filter is on yours.

For instance…

• If you’re a writer, the test is whether you can describe the piece you’d write so completely that AI can draft it and you only edit.

• If you’re a designer, it’s whether you can describe the layout, the constraints, the brand voice, the user need, in enough specificity that the model produces something you’d actually ship.

• If you’re a consultant, it’s whether you can encode your judgment into a prompt that another person could run without you in the room.

I see this recognition starting to surface, even when people can’t quite name it yet.

Out of all the workshops I run, the most-requested format is the same every time:

“I just want to watch you build live and explain why you’re doing things.”

People who’ve worked with AI long enough know there’s something deeper going on than just better prompts. They want to see how I think.

All the attention is on the machines. Better prompts, better tools, better models. The gap is somewhere else entirely. It’s THE gap that separates the top 5-10% of AI users from everyone else.

This is the new core skill and what people want to watch when I run a workshop.

Directing.

How directing actually works

Film directors hold a vision, communicate it, watch what gets made, and call cut when it misses. With AI, it’s the same skill. Knowing what you want clearly enough to describe it. Recognizing when the output drifted. Closing the gap, in real time. The skill that reads as “soft” until you watch someone good at it produce ten times the output of someone with twice the technical training.

Every director develops a signature over time, the moves they reach for and the ones they skip. You already have one. It’s in every prompt you’ve written. You just haven’t read them as a body of work.

I want to give you a receipt for this, because I am the receipt.

Six months ago I had never written a line of code. I presented at a hackathon. I still haven’t written a line of code. I’ve shipped more than fifty custom Claude skills since then, run an autonomous agent system on a Mac Mini in my living room, and ghostwrite a weekly newsletter for a marketing legend whose voice I extracted into a system. None of that required me to learn syntax. All of it required me to learn how to specify, diagnose, and direct.

The signature got built one directed session at a time.

The clearest version of this lives inside one of my retainer clients. Three people on the team I work with closely, three completely different AI installs.

1. Tom runs sales. He got a system that finds, ranks, and pre-qualifies leads, generates custom outreach, and plugs straight into his email and CRM.

2. David handles client delivery. His install is a Claude Project loaded with messaging for three different avatars, strategy, and deliverables. It’s wired into the Project Management System (PMS), the calendar, and Fireflies, so every meeting he runs gets processed automatically. Tasks land in the PMS. Client context lands in Google Drive. Claude proactively suggests the next step on whatever deliverable is open.

3. Sarah runs hiring. Her system sorts the resume inbox, scores candidates against a company rubric, and generates a personal profile card on each person for the leadership team to review.

The reason all three worked isn’t that I knew Python (I absolutely do not). It’s that I sat with each of them long enough to specify what their work actually was, where the friction lived, and what shape an AI assist needed to take to slot in without breaking what already worked. The build was the easy part. The directing was the whole thing.

It’s probably not a coincidence this is where I am focusing. The two things I spend most of my time on are this (AI implementation) and Cognitive Fingerprint™, which is extracting what makes someone unique and articulating it back to them clearly enough that they can use it.

Both are articulation skills and directly fall under…directing. Which is why the test the new interview is filtering for isn’t a tech skill at all.

What this means for you

You also probably suspect, if you’re honest, that the muscle for the new test is underdeveloped. You can prompt. You’ve used the tools. You’ve watched a few people on YouTube spec out a Claude Code build and thought “I could do that,” and then tried it on something real and noticed the model drifted, you didn’t catch it, you cleaned up the output by hand, and you wondered if you’d actually saved any time.

You’re not missing a tool.
You’re not missing a better model.
You’re missing reps at directing.

Reps come from running real work through repeatable patterns.

Knowing that directing is the skill, and seeing your own directing on the page, are two different things.

The AI Director’s Script is the kit. Bring three real prompts you’ve used recently in AI sessions. The kit walks you through each prompt the way a director reads a draft: for vision, clarity, drift, and cut. Then rewrites them as scripts that work without you in the room. By the end, you’ll have three rewritten prompts and your directing signature named: the move you reach for, the move you skip.

Here’s how it works.

🎬 The AI Director’s Script

Every great film starts with a script. Every great AI session does too. This is yours.

How to use this kit

Bring three real prompts. Use whatever’s actually in your AI history this week. The email draft, the analysis request, the plan you sketched, whatever rough thing you typed last. The rougher, the better.

Run the prompt below in any AI assistant. Claude, ChatGPT, Gemini all work, anything that holds a multi-turn conversation. The session takes 15-20 minutes if you stay honest with it.

Expect to be asked what you actually meant. The kit pushes back on vague answers. That pushback is the directing rep happening in real time. The discomfort is the work. If you breeze through, your prompts were probably already in good shape, and you’ll have the rewrites to prove it. If you get stuck, that’s the gap you came here to find.

You’ll walk away with three things. Three rewrites you can paste into your next AI session immediately. Script notes on each prompt, naming what was thin and what got fixed. And your directing signature: the move you reach for, the move you skip, and the habit change that closes the gap.

What it looks like in practice

For example, let’s run the kit on a simple prompt. Take one almost everyone has typed at some point:

“Write me a LinkedIn post about how AI is changing knowledge work.”

The kit walks you through what that prompt is missing.

• Vision (what you actually wanted): the prompt doesn’t say. “Thoughtful, not cliché” might be in your head, but the model can’t see it. Without vision, you get the AI default. Generic bullet points about productivity gains.

• Clarity (could anyone else have written this prompt?): no audience, no length, no format, no examples of good. A stranger picking up this prompt would produce exactly the bullet-point post you didn’t want.

• Watch (where did the output drift?): the whole output was the drift. You wanted a personal angle. You got “automation of routine tasks” and “new opportunities for growth.”

• Cut (what would have prevented it?): a sentence naming the audience. A sentence naming what the post should NOT do. A sentence about what shape “good” takes.

The rewrite:

“Write a LinkedIn post for marketing peers who use AI daily but don’t think deeply about it. 200 words, first-person, conversational. Open with one specific moment from this week where I noticed AI changing how I work. Skip the productivity gains and automation framing. Something subtler. No bullets, no buzzwords, no ‘in today’s fast-paced world.’ End with a question that invites disagreement in the comments.”

Your directing signature:

You reach for vision well. You know what you don’t want, even when you can’t name what you do want. You skip clarity hardest. Your prompts assume the model knows your audience and your taste. The habit change: name the audience and one anti-pattern at the start of every prompt. One thing the output should NOT do.

Read more

If you’re subscribed to Signal > Noise, surprise, it’s about to be called Cognitive Fingerprint™.

Probably not a big surprise and what this should have been from the start. But like everything, it’s part of the evolution.

But here’s a bit more on the “why.”

Every conversation I see right now treats humans like line items to be replaced.

• Employees as costs

• Headcount as a number to shrink

• AI as the lever

I want to spend my time on what makes you, specifically, you.

The invisible patterns that separate your best work from anyone else’s. The stuff that gets ignored the second someone hands you a generic playbook. AI doesn’t replace that. Pointed at the right thing, it amplifies it.

My goal is to show you how to point it and walk you through the people whose patterns got sharper the second AI showed up.

That’s what CF has always actually been. Now the newsletter has the right name.

Expect a lot more from here and other places. My three partners, , , and , are going to show up in their own unique ways more often too. They’re a big part of why this is working.

Anyway. Dr. Pepper. Keep reading.

— Max

Subscribe now

I was watching baseball this weekend when a commercial came on for Dr Pepper. Instead of getting another snack, I found myself asking Perplexity a question my marketing brain couldn’t get past.

Is Dr Pepper really made with 23 flavors, or is that just a gimmick?

The short answer... yes…but it’s complicated. The formula uses 23 flavor compounds and parent company Keurig Dr Pepper has confirmed it.

But here is where it get’s fun. Nobody outside the company knows what they are.

I’m not talking about the ingredients on the label (carbonated water, corn syrup, caramel color, phosphoric acid, caffeine, sodium benzoate, and a catch-all bucket called “natural and artificial flavors”).

Or any of the viral leaked lists floating around since 2008 with almond and plum and a Bob Hope joke about prune juice (“a drink that moves you!”) That’s all fan fiction. The actual 23 are sealed in a proprietary blend the FDA lets companies hide.

Dr Pepper’s actual recipe is split across two separate Dallas bank vaults. Half in each, like two pieces of a doomsday riddle. 141 years of refusing to show you what’s in the bottle.

And this is where I had to keep digging.

Four brands. Each with a secret vault.

OK, there has to be other companies who have used this trick before, right? Turns out, quite a few:

• Coca-Cola calls their formula “Merchandise 7X,” and they keep it in a vault at the World of Coca-Cola museum in Atlanta. Only two executives know the recipe, and they’re forbidden from flying on the same plane.

  

• KFC splits the 11 herbs and spices across two separate factories. Neither factory has the complete formula. The final mix happens at a third site with 24-hour surveillance.

  

• Chick-fil-A keeps the original sandwich recipe in a literal safe at their Atlanta headquarters. They won’t tell anyone who holds the combination.

But what’s crazy (or so I thought) is that none of them have patented their secrets. Every instinct says you’d want that protection. Patents are how businesses are supposed to guard their inventions.

I kept clicking because something about it felt backward until I read the fine print.

Patents expire after 20 years and require full public disclosure upfront. You show the world your formula in exchange for 20 years of protection, and after that anyone can copy it. Trade secrets, on the other hand, last forever... as long as you never show anyone what’s in the bottle.

The mystery is the asset. Locking it up is the business model.

It all made sense when I found the 1963 Dr Pepper lawsuit. At the time, a federal anti-monopoly rule required major bottlers to carry non-competitive soft drinks in their facilities. If you were Pepsi, you had to bottle anything that wasn’t a cola. Dr Pepper’s lawyers spotted the gap and sued the federal government for the right to be legally classified as a non-cola. “A pepper type drink,” they argued.

They won and because Pepsi had no competing pepper-type drink, the ruling forced Pepsi’s bottling facilities to distribute Dr Pepper nationwide.

A category of one became a legal category of one, which became a national distribution line.

That’s when it stopped being about soda for me.

I work with consultants, coaches and experts for a living. Their whole business is proprietary thinking they’ve built over years of work. So when I saw a 141-year-old soda company still guarding its formula, I had one question... why don’t the experts I work with have a formula to guard in the first place?

• They have the knowledge

• They have decades of experience

• They have something proprietary buried in there somewhere

They’ve just never named it, shaped it, or locked it up.

The soda companies figured out the move in the 1880s. The experts I work with still call it “how I do things.”

I mean, look how powerful this concept is!

141 years after they locked up the recipe, strangers on YouTube are still making free content trying to guess what’s in it.

The vault does the marketing too.

“Take two parts Coke, one part root beer, drink it and tell me that isn’t Dr. Pepper.”

That video is somebody’s spare-time conspiracy theory. Dr Pepper didn’t pay for it.

In 2009, a Tulsa man named Bill Waters was poking through a Texas Panhandle antique store. Under a wooden medicine crate, for $200, he found a tattered ledger from the original Waco drugstore where Dr Pepper was first served in 1885.

Inside, a handwritten recipe titled “D Peppers Pepsin Bitters.” Heritage Auction Galleries put it up for auction expecting $50,000 to $75,000. The AP ran the story. Dr Pepper’s parent company had to issue a corporate statement saying no, that isn’t the formula. The book didn’t sell. The coverage ran anyway.

The vault just keeps generating it.

Now look at yourself.

You’re not a soda. You’re an expert. You sell thinking.

So if you read that story, nod, close the tab, and go back to writing another LinkedIn post about “the top ten ways to do x,”

That’s the problem I want to talk about.

Every expert I know has a vault.

They just don’t know it’s there.

The Legal Pad Test

Early in his career, marketer Todd Brown sat in a coffee shop across from a prospect. A legal pad on the table. Five columns, one per coach. Four rows of criteria. Price. Credentials. Testimonials. Guarantees. Todd was column three, sandwiched between two coaches with nearly identical resumes.

The prospect looked up and said, “You all seem pretty similar. I guess it just comes down to price and who I like best.”

Todd was being priced like canned corn.

He calls his framework the Unique Mechanism Matrix. It plots two axes, uniqueness and substance, across four quadrants.

Commodity Hell is the default address. Nothing unique, no substance behind the differentiation. Price war territory.

Some climb to Commodity Plus. Good at the work, forgettable. The prospect can’t describe what you do that the next person doesn’t.

A dangerous few land in Faux Mechanism. Named something that sounds proprietary. “The 7-Step Blueprint.” “The Accelerator Method.” Press on it and the substance dissolves.

Few reach Unique Mechanism. Unique AND substantive. Category of one. The prospect stops comparing you cell by cell because there’s nothing to compare you to.

The spreadsheet only works when everyone fits in the same columns.

Three Vaults. Two You Can Open.

When you hear “Dr Pepper has 23 flavors” or “Chick-fil-A has a safe,” you translate it as “I need a patent.” You don’t have one. The whole thing feels out of reach.

Todd’s move is to break the mechanism into three types. One isn’t available to you. The other two are wide open.

• Product Mechanism is the what. The ingredient, the formula, the molecule. Dr Pepper’s 23. Coke’s 7X. You can’t have this. Skip it. You’re stealing the move, not the molecule.

• Process Mechanism is the how. The sequence, the method, the protocol. KFC splits the blend across three factories so none has the complete recipe. That’s a Process. The specific way you take a client from point A to point B that no one else is doing in that exact order. Available to you.

• Philosophical Mechanism is the why you. The contrarian position. The belief others in your industry don’t hold. Chick-fil-A closes on Sundays when every other major chain is open. That’s a Philosophy. The thing you argue at dinner that other people in your field don’t argue. Available to you too.

Every expert has both. Few have named either.

The Paper Tiger Test

Before you name anything, a warning.

Naming is the tip. The mechanism lives underneath.

Todd calls the failure mode the Faux Mechanism. A paper tiger that looks real from the front and tears apart the first time someone presses on it.

The test is one question. “Why does your approach actually work?”

• A faux mechanism answers with vibes. “It’s a proprietary system.” “It just works.” “Clients love it.”

• A real mechanism answers with specificity. When we do X in step 2, it surfaces Y the client wasn’t conscious of, which means the work in step 4 lands on actual constraints instead of performed ones.

If you can’t answer “why” with reasons, you don’t have a mechanism. You have a slide deck.

A Name That Dies in a Calendar

The second mistake is a name that dies in obscurity.

Great names pass five filters. Brown gives one rule at the core. A mechanism name should sound like a thing in a textbook, not like sales copy. The other four are how I make sure that one actually holds.

1. It sounds like a thing. Put a ™ after it. Does it look natural, or does it look like you’re dressing up a phrase?

2. It creates curiosity. The person hearing it asks “what’s that?” instead of “cool, thanks.”

3. It’s memorable. One to four words. The person can repeat it at dinner after hearing it once.

4. It signals the mechanism, not the benefit. The name should hint at how it works. The move, the lever, the cause. “Lumbar Reactivation” tells you the muscle group and the action. “Cash Flow Compounding” tells you the engine. “The Back Pain Relief Protocol” tells you the wish, which is product copy. Save the benefit-driven name for the wrapper, not the mechanism inside.

5. It’s differentiated. Doesn’t sound like the 12 other frameworks in your industry.

The named methods I see fail three out of five. “The Growth Accelerator Blueprint.” Sounds proprietary, carries no information, fails the cocktail party test, blends into every other accelerator blueprint out there.

A mechanism without a name dies in someone’s calendar.

Your Move This Week

Not “find your unique mechanism.” Too abstract. I’ve done it. Forty minutes of staring and getting distracted by YouTube.

Instead, pick one specific thing you do with a client that no one else in your industry is doing. One. Be boring about it. Maybe it’s a question you always ask on call three. Maybe it’s a template you hand them before the kickoff. Maybe it’s a specific sequence you use to run a teardown.

Now name the thing. Four words or less and run it through four checks.

• Name. Does it pass the five filters? ™-ready, creates curiosity, 1-4 words, signals the mechanism (not the benefit), different from the 12 other frameworks in your industry.

• Type. Is it a process (the how) or a philosophy (the why you)? Be honest. If it’s “kind of both, a little of each,” it’s probably neither yet.

• Substance. Can you write one paragraph that answers “why does this actually work?” without using the words “experience,” “proven,” or “proprietary”?

• Vault. Can you talk about the concept publicly while keeping the specific sequence private? If the whole thing lives on your sales page, you don’t have a vault. You have a brochure.

If any one of those breaks, you’re not done. Keep working on the thing itself.

If all four hold, you just moved out of column three.

The Gap: The Value Builder

You have the matrix, the three types, the faux tiger test, and the five filters. You have the shape.

Your mechanism is buried in years of client work, call notes, and case studies. The same invisibility that makes it proprietary makes it hard to see alone. The chef can’t taste their own kitchen.

The Vault Builder is three prompts that run this framework on your own material. Feed it your case studies and call transcripts.

• Prompt 1 surfaces 5-7 candidate mechanisms you actually practice (not the ones you claim on your website).

• Prompt 2 runs each candidate through a five-round collapse test so you know which ones hold up under pressure.

• Prompt 3 generates 20 name candidates scored against Todd’s five filters and recommends the top three.

Two formats ship together. Copy-paste prompts you can run in any Claude chat. Or a downloadable Claude skill (vault-builder.skill) hosted at cognitivefingerprint.ai/prompts. Installs once. Runs the whole protocol with a single command. Use whichever fits your setup.

What took me months of Cognitive Fingerprint development, compressed into an afternoon with material you already have.

Read more

Then three things happened in 90 seconds:

• A regional jet missed its speed restriction

• A FedEx heavy requested an early turn

• A Cessna called in on the wrong frequency

The controller didn’t panic. He didn’t freeze. He looked at his radar strips. Moved two of them to the left side of his bay. Keyed the mic. Started sequencing. One instruction at a time. No rush in the voice.

But here’s what he didn’t do.

He didn’t try to hold all twelve aircraft in his head.

Every approach controller at O’Hare manages 20 to 25 aircraft at a time. That’s 857,000 movements a year. One every 37 seconds. The system works because controllers externalize almost everything.

Each aircraft gets a paper strip. Callsign, type, altitude, assigned heading. The strip sits in a physical bay, and when the controller issues an instruction, they mark it. The radar screen shows positions. The strips hold the plan.

The controller’s working memory holds almost nothing. That’s by design.

On January 29, 2025, a controller at Reagan National was working a quieter shift. But the traffic was mixed in a way the system wasn’t built for. An Army helicopter was flying a training route along the Potomac. At the same time, a regional jet was circling to land on a runway whose approach path crossed the helicopter’s route. The two aircraft were on different radio frequencies which meant neither crew could hear the other.

At O’Hare, two separate controllers would have handled this. At Reagan that night, one controller was managing both. Federal investigators found the combined workload caused a loss of situational awareness. The controller lost track of the helicopter. In a disaster that years of warnings should have prevented, sixty-seven people died.

The Reagan controller’s working memory hit a wall. It wasn’t a skills wall or a training wall. It was a biological one, baked into every human brain, unchanged since before we had language to describe it.

Two types of traffic converging on the same airspace, two radio frequencies to monitor, and no second controller to split the load.

His brain was asked to hold what the system should have carried. It couldn’t.

Subscribe now

Before You’ve Done a Single Thing

Your thumb finds your phone before your brain finds the morning. A red notification badge. An email you shouldn’t have opened. A calendar reminder for something you forgot to prepare. This list is already overflowing.

The client request you’ve been avoiding. Something about taxes. Your kid’s school project. The 30 emails you haven’t opened. A promise you made to review someone’s deck. A dentist appointment you keep rescheduling. And a nagging feeling you’re forgetting something important that you can’t quite remember.

Each one of those is an aircraft on your radar. And you’re holding all of them in your head.

And if you are like me there are no strips and no real system.

Just you, staring at the screen, wondering why you feel paralyzed before you’ve done a single thing.

What’s Actually Happening in Your Head

“The Magical Number Seven, Plus or Minus Two” (1956)

George Miller figured out the first piece in 1956. His paper, “The Magical Number Seven, Plus or Minus Two,” pinned down a hard limit. Human working memory holds about seven chunks of information. Not items. Chunks. The ceiling is the same either way.

(Nelson Cowan revised it downward in 2001. His research suggests the real number is closer to four. That’s it. That’s what you have to work with.)

Miller himself was almost apologetic about it. He opened the paper by saying he’d been “persecuted by an integer.” The number is a hard constraint. Like the number of plates a waiter can carry. You can get better at stacking them, but you can’t change the size of your hands or arms or shoulders. (Some of these waiters are insane with how many plates of sizzling fajitas they can stack just on their forearms.)

“On Finished and Unfinished Tasks” (1927)

The second piece came earlier, from a Berlin restaurant in the 1920s. Kurt Lewin, a psychologist, noticed something about a waiter. The man could remember every detail of every open order... who ordered what, which modifications, which table. But the moment the bill was paid, the order vanished from his memory. Gone.

Lewin’s student, Bluma Zeigarnik, turned that observation into an experiment in 1927. She gave people tasks. Interrupted some of them halfway through. Then tested what they remembered. The interrupted tasks were recalled roughly twice as often as the completed ones.

Your brain does not let go of unfinished work. It keeps pinging you over and over like a social media app with those annoying notifications.

Every open loop takes a slot and holds it. Every vague commitment. Every “I should really...”

But oh buddy, it gets worse. The vaguer the task, the harder your brain grips it. “Something about taxes” can’t be chunked. Can’t be filed. It just sits there, taking up space, because your brain doesn’t know what to do with it and refuses to let go until it does.

I watched this happen to a client last month. Brilliant marketing guy about to launch a new campaign. He popped into our Zoom session and I could see the hopelessness on his face right away. He said, “I can’t think straight and I don’t know why.” We spent 11 minutes writing down everything competing for his attention. The list hit 23 items. He stared at it and said, “I thought it was three or four big things.” It wasn’t. It was 23 invisible things. Most of them small. None of them planned. All of them taking a slot.

You’re not carrying 14 tasks. You’re carrying 14 open notifications that your brain can’t swipe away.

So it feels like you’re overwhelmed by how much you have to do. Like the volume is the problem.

In reality, you never had too much to do. You just couldn’t see what you were carrying. The tasks aren’t hard. They’re hiding. And you can’t take action on what you can’t see.

“Consider It Done” (2011)

The third piece came from Roy Baumeister in 2011. His paper was called “Consider It Done.” The title tells you everything.

You don’t have to DO the task to release it from working memory.

You just have to make a plan.

A specific plan. Not “I’ll deal with it later.” A plan with a when and a how. “I’ll check the tax deadline tomorrow at 9 AM on the IRS website.” That’s enough. Baumeister showed that once the brain has a concrete plan, it files the task as handled. The intrusive thoughts stop. The notification clears.

The controller at O’Hare doesn’t remember which planes are descending. The strip remembers. His brain is free to handle what’s happening now.

Your brain is trying to be the strip AND the controller. At the same time. For 14 aircraft.

No wonder you’re anxious before 9 AM.

The People Who Look Calm

The people you know who seem calm under heavy workloads aren’t carrying less than you. They’re not better at tolerating the weight. They’re better at putting it down.

Their working memory is nearly empty. They probably have more to do than you. But nothing is living in their head unplanned. Every open item is visible somewhere. Every item has a next step. Every next step has a slot.

The anxiety doesn’t come from the work. It comes from it flying out of control in your head.

And the most counterintuitive part... the anxiety drops before you do a single task. Baumeister proved it. The moment the items move from your head to a visible, sequenced list, something loosens. Your brain stops clenching around every open item and lets the list hold them instead.

You don’t need to do less. You need a radar.

The Radar

Ten years ago, your scope held fewer aircraft. Now Gloria Mark at UC Irvine finds the average person switches context over 400 times a day. Each switch burns a slot you don’t get back before the next one lands. You’re running 2026 traffic volume on a 1956 cognitive architecture.

The next time the blob hits, try this. One thing. Five minutes.

Open a blank page. Write down every item competing for your attention until the page stops surprising you. Work, personal, half-formed, vague, embarrassing... all of it. Don’t organize. Don’t prioritize. Just get it out of your head and onto something you can see.

Then... before you do a single one... write the next action next to each. Not the project. Not the goal. The action. “Something about taxes” becomes “Google quarterly tax deadlines.” “That thing for Jake” becomes “Text Jake and ask for latest deck.” If you can’t name the action in one sentence, the item is still vague. Sharpen it until your brain can chunk it.

That’s the moment the pressure drops. Not when you finish the list. When your brain believes every item has a plan.

Baumeister proved it. Your brain treats “I know when and how” almost identically to “done.” The strips are in the bay. The radar is on. You can start working now.

One page. One pass. The anxiety has nowhere left to hide.

The Article Is the Proof

Your brain has been running an approach control facility with no external systems this entire time. Every unfinished thought from this morning, every half-read email, every “I should really...” has been sitting in working memory. Taking up one of your four slots. Refusing to leave.

The most anxious people I know aren’t the ones with the most to do. They’re the ones with the least visibility into what they’re carrying. The blob isn’t big. It’s opaque.

Feels like a character flaw. Like everyone else handles this better. Like you should be able to just... manage it.

The cost is that you’ve been blaming yourself for a design problem. Your working memory was never meant to carry your entire life.

• Miller told you in 1956

• Zeigarnik told you in 1927

• Baumeister told you the fix in 2011

The anxiety is a signal. But your system has no radar.

Build the radar. The anxiety has nowhere left to live.

What It Looks Like When the Radar Works

Marcus is 42. Solo leadership coach. Recently separated. Six active clients, keeps dropping balls. He sat down on a Monday morning with coffee feeling paralyzed.

He opened an AI conversation and started talking.

Marcus: I have a client deliverable due Friday, Rachel, she’s waiting on a leadership assessment report. And I need to prep for a call tomorrow with my newest client, James, and I haven’t even looked at his intake form yet. There’s invoicing I’ve been avoiding for like two weeks. And I need to figure out the kids’ schedule for next month because Sandra and I are still fighting about who has them when and I just…I can’t think about that right now but it’s sitting there...

The AI didn’t advise. Didn’t organize. Just said “keep going” and “what else?” When he said “it’s all a blob,” it said “That’s exactly why we’re doing this. Keep going.”

He kept going. Therapist he’s been meaning to find for two months. Conference early bird he keeps missing. Newsletter he skipped last week. Emails he’s afraid to open.

Then it mirrored everything back.

AI: That’s 15 items. They’re all here now. You can stop holding them.

Marcus: Seeing it like that is... weird. It felt like way more in my head.

Then it clarified each item. One question at a time.

AI: The kids’ schedule with Sandra — is there an action on your side?

Marcus: I need to text her with my availability for the next six weeks. I keep not doing it because every conversation turns into a fight. But the action is just sending the text.

AI: The therapist — do you have anyone in mind, or is step one finding names?

Marcus: I could ask my doctor for a referral I guess.

AI: So the action is: text or call your doctor’s office and ask for a referral.

Marcus: That’s so obvious. Why haven’t I done that.

Then it built his day.

Today’s queue:

1. Email scan (~25 min)

2. James call prep (~1 hr)

3. Tom — reschedule text (~5 min)

4. Invoicing (~20 min)

5. Rachel’s report (3–4 hrs) — This is the win.

You have 5 actions today, 7 this week, 2 parked, 2 waiting. Your next move is #1. Open the inbox. Go.

Marcus: I feel weirdly better. Like nothing changed but I can see it now.

Nothing changed. He just stopped holding all of it in his head at once.

Five minutes. Fifteen items. A sequenced plan. The anxiety dropped before he touched the first task.

The Gap

You now know what’s happening in your head. You have the blank page method. And you just watched it work on someone carrying 15 invisible items.

But here’s what I’ve watched happen next. You’ll try the blank page once. Maybe twice. The third time the blob hits... you’ll open the page, write four things, get stuck on the fifth because it’s vague, feel the anxiety spike again, and close the document.

The method works. The follow-through doesn’t.

Because doing this yourself requires you to be your own interviewer, your own clarifier, and your own sequencer... while your working memory is already full. You’re asking the overloaded controller to also be the strip, the bay, and the radar. At the same time.

Marcus didn’t do the blank page. He had a conversation. The AI asked the questions his brain couldn’t form. It held the items his working memory couldn’t carry. It clarified what he was too foggy to sharpen on his own. And it built the queue he didn’t have the bandwidth to sequence.

That’s the difference. The insight is free. The mechanism that runs it while your brain is drowning is what I built next.

The Clarity Protocol is a set of AI prompts you paste into Claude, ChatGPT, or whatever you already use. They do what Marcus experienced... interview you, surface everything you’re carrying, clarify each item into a concrete next action, and hand you a sequenced plan. Three prompts, a voice mode version that does it all in one conversation, and the full walkthrough are below.

Build the radar. The anxiety has nowhere left to live.

The Clarity Protocol

Four prompts. Paste any of them into Claude, ChatGPT, or Gemini. They work the same way Marcus experienced it….

Read more

This article first appeared in the “Beware the Default” newsletter but I really enjoyed writing it and I think there are some good reframes in here that you will enjoy reading.

In 1925, the average American woman spent five and a half hours a week on laundry. She was scrubbing on a washboard. Wringing by hand and hanging everything on a line and praying it didn’t rain.

By the 1950s, the automatic washing machine had reached most American homes. The promise seemed obvious: all that scrubbing and wringing, gone. Free time would appear.

But Ruth Schwartz Cowan studied what actually happened, drawing on time-use research by sociologist Joann Vanek. The finding was not what anyone expected.

By the 1970s, women were spending more time on laundry than before the machines arrived.

Total housework had gone up, not down.

The machines made laundry easier, so people washed clothes more often.

Sheets that used to last a week now lasted three days. Standards for what counted as “clean” quietly recalibrated upward. The stain your grandmother would’ve ignored became the stain your mother couldn’t tolerate.

Cowan called it the “great irony of household technology”. The machines didn’t free anyone.

They moved the bar.

I know, I know…You already see where this is going.

The AI productivity story runs on the same assumption the washing machine did.

Do the same work faster, get the leftover time back. That’s the pitch. That’s the implicit promise every time someone posts a “saved me 4 hours” screenshot.

But the time doesn’t come back. It gets absorbed.

Subscribe now

How the Bar Moves Without You Noticing

Aruna Ranganathan and Xingqi Maggie Ye spent eight months inside a 200-person tech company watching this happen. Their study, published in Harvard Business Review this February, tracked 40 workers who adopted AI tools.

The findings should’ve been a celebration. Workers got faster. WAY faster.

Then three things happened, in order:

First, scope expanded. Workers who finished tasks faster didn’t log off early. They took on more work. The freed-up time just... filled. The same way a new lane on a highway fills with traffic within months like the extra space never even existed.

The capacity didn’t create slack. It created demand.

Nobody sent a memo. The work just appeared. And because it appeared at the exact moment they felt productive, they called it ambition.

It was absorption.

Second, hours extended. Workers started sending prompts and requests after hours. The tool was always available, so the work became always available. The boundary between “work time” and “free time” didn’t blur. It dissolved. Email did this exact thing 20 years ago. Office workers now spend 28% of their week managing email... a technology that was supposed to make communication faster.

But, as you can see (and probably know), the pattern is the same. Lower the friction of doing a thing, and people do more of the thing. Not less. The bar was moving. Everyone called it dedication.

Third, quality dropped. The cognitive load of broader scope plus longer hours plus constant output evaluation added up. The researchers’ conclusion: AI didn’t reduce work. It intensified it. It felt like a mystery. It was the bill coming due.

The gain was the story people told on the way in. What they found was a trap.

Speed up the work -> Absorb the speed into expectations -> Extend the hours to meet the new expectations -> Burn out under the weight of a bar that moved while nobody was watching.

We are watching the washing machine cycle get compressed from decades into months.

The Part That Actually Costs You Something

There’s a layer underneath the burnout that’s harder to measure but worse to lose.

A team led by Elena Hayoung Lee published a study in Nature this March. They gave 269 people tasks to complete with AI assistance. The variable was:

• One group received AI outputs passively (the AI did the work, they reviewed it).

• The other group collaborated actively (working with the AI, shaping the output together).

The passive group finished faster. They also reported lower self-efficacy, lower sense of ownership, and less meaning from the work.

Sit with that for a second. The people who let AI handle it... felt less capable afterward. Not because the AI failed. Because they succeeded without being involved.

The washing machine took away the physical labor of scrubbing. It also took away the thing that made a clean shirt feel earned. The same exchange is happening in your work right now.

You used to write the first draft. Now you edit the AI’s first draft. The output might be better. The feeling of authorship is different. And that difference goes deeper than feeling. It shows up in how you talk about the work, how much you’ll fight for it in a meeting, how deeply you understand what you shipped.

Ownership is a signal of understanding. When it fades, the understanding goes with it.

And honestly, it’s something you have to experience firsthand to really get it. One AI-generated article is easy to dismiss. One delegated task, one social post you didn’t really write. That’s nothing.

The feeling comes later. When it becomes a habit. When you look up one day and realize you feel disconnected from your own work. Not tired. Disconnected. Like the thing you’re producing stopped being yours and just became more noise.

Until I wrote this, I could feel that. I just couldn’t name it.

Dario Amodei, CEO of Anthropic, raised the same question on the Lex Fridman podcast. In a world of increasingly powerful AI, where does meaning come from? His answer wasn’t complicated. Work is a source of deep meaning for most of us, he said, and that won’t change no matter how capable the tools get. The need for meaning doesn’t get automated away.

The Lee study already told you the answer. The people who felt less capable afterward handed the work over and reviewed the output. The ones who kept their sense of ownership stayed inside the thinking as it happened. Shaping it. Treating the AI as a collaborator.

That’s the variable. Stay inside the work or step outside it.

Use AI to move faster through the parts that don’t require you. Stay inside the parts that do.

Why This Time Is Faster

The washing machine took 50 years to fully recalibrate expectations around cleanliness. Email took about 15 to become the thing it was supposed to replace (meetings, memos, phone calls) plus a dozen things it was never meant to be.

AI is doing it in months.

The cycle is the same. The clock is faster. Your organization adopted AI tools sometime in the last year. The expectations have already moved. The “normal output” for a Tuesday has already been redefined by whoever on your team figured out Claude first. And once one person’s output resets the benchmark, everyone else is running to match a bar they didn’t see move.

The Question Worth Sitting With

Before you fill the time AI freed up, ask who decided it needed filling.

Not “how do I use this time productively.” That question has already been colonized by the same expectation inflation that turned the washing machine into more laundry.

Just: who decided?

If the answer is “my calendar filled itself” or “there was more work so I took it” or “I guess I just... did”... you’re living the washing machine paradox. The time was never yours. The bar moved, and you moved with it, and the gap between where you were and where you are now has been papered over by the feeling of being productive.

Productivity is the feeling. Leverage is the question.

The difference between the two is whether you chose what to do with the hour you got back, or whether the hour got taken before you noticed it existed.

Your Organization (Or Client) Is Reading This Too

Your ego will tell you this is a personal discipline problem. That you just need better boundaries, better time management, a system.

But Cowan’s research showed something the self-help version misses. The washing machine didn’t just change individual behavior. It changed what society expected of a household. The standard shifted, and every individual decision happened inside that shifted standard. You couldn’t opt out of “clean” being redefined.

You could only decide how to respond.

The AI version of this is already in motion. The company that saw one team ship 3x faster with AI didn’t celebrate and go home. It asked why the other teams weren’t doing the same. The bar moved for everyone because it moved for someone.

You’re not going to solve this with a morning routine.

The Oldest Trick in the Technology Book

In 1983, when Cowan published her research, the response was predictable. People said it didn’t apply to them. Their washing machine really did save time. They really were spending less effort on laundry than their mothers.

They were wrong, and they were wrong in the exact way that made the data invisible.

The standard moved slowly enough that each incremental shift felt like a personal choice rather than a structural one.

The AI bar is moving faster. Which means you’ll notice it sooner...if you’re looking.

But most people aren’t looking. They’re too busy filling the time they saved.

-Max

This is what we talk about inside Beware the Default. Not productivity tips. The invisible assumptions underneath them. If that’s your kind of room, we’d like to have you.

Sources

• Cowan, R.S. (1983). More Work for Mother: The Ironies of Household Technology from the Open Hearth to the Microwave. Basic Books.

• Ranganathan, A. & Ye, X.M. (2026). “AI Doesn’t Reduce Work—It Intensifies It.” Harvard Business Review, Feb 9, 2026.

• Lee, E.H., Yin, Y., Jia, N., & Wakslak, C.J. (2026). “Relying on AI at work reduces self-efficacy, ownership, and meaning while active collaboration mitigates the effects.” Scientific Reports (Nature), March 15, 2026.