Knowledge work just got a public funeral by Anthropic

Don't learn AI, but build and play with AI to get outcomes

This is a FREE 🏴‍☠️ Founding Members 🏴‍☠️ post. Founding Members get access to the Pirate Eddie Bot to ask category design questions, weekly actionable insights, the full library with 30+ audiobooks 200+ mini-books, and more. See the Founders Deck here.

---

Dear Friend, Subscriber, and Category Pirate,

Last week, a spider chart from Anthropic’s new labor market impact report made waves. It shows two things: the blue area is what AI can theoretically do across every job category. The red area is what people are actually using it for.

Andrew Curran@AndrewCurran_

Striking image from the new Anthropic labor market impact report.

8:27 PM · Mar 5, 2026 · 7.12M Views

---

562 Replies · 2.28K Reposts · 13.6K Likes

Look at the gap.

In business and finance, computer science, architecture, engineering, management, education, and legal, AI can theoretically handle the vast majority of tasks. But actual usage? A tiny red sliver in the center.

That gap is the entire Creator Capitalist thesis in one image.

Remember, you don’t have to outrun the bear, just the slowest person?

• You don’t have to master AI at the blue level.

• You just need to be ahead of the red.

• Red is dead.

And the people moving toward the blue? We call them Creator Capitalists.

“Learning AI” Leaves You in The Red

Learning AI is a legacy Knowledge Worker’s response.

It treats AI as one of many skills, books, or classes to add to your resume. The problem is that AI is happening in real time, and knowledge takes months or years to be codified in a class, book, or some other form of legacy expertise.

By the time you find an AI expert, that expertise has expired.

And you will be far behind.

When Category Pirates started its AI journey over a year ago, we were told by an expert that we had to:

• Upload documents, quizzing, and correcting ChatGPT for 4-6 hours per day

• Repeat 7 days a week for 6 months

This seemed really daunting and demoralizing to start!

It turns out the advice was wrong.

Was the expert getting great results? Yes. Was it because of the hundreds of hours of training? Not likely, given the memory limitations of LLMs at that time.

It was far more likely a correlation than a causation. OpenAI was shipping its most aggressive model improvement run in history at the same time.

• September 2024: o1-preview launched with chain-of-thought reasoning roughly 8× better at complex tasks.

• November 2024: Projects and persistent Memory launched, making the manual daily uploads redundant.

• December 2024: o1 full release — 89th percentile on Codeforces, PhD-level accuracy on science benchmarks.

• February 2025: GPT-4.5 cut hallucinations by 40% (61.8% → 37.1%) and improved factual accuracy by 4× over GPT-4o on standardized knowledge tests.

This meant that a beginner was just as well-positioned as an expert.

You are far better off building as a beginner than learning from an expert. That's the Creator Capitalist approach to AI—and it's one of the reasons we devoted two full chapters of the book to it.

‘Build with AI’ in the Near Blue

19-year-old Zach Yadegari had just been rejected by his dream colleges like Stanford, MIT, Harvard, Columbia, Princeton, Duke, and Cornell. Despite his 4.0 GPA and 34 ACT score.

And yet, he just sold Cal AI, his $30 million ARR photo-based calorie tracking app, to MyFitnessPal for $50 million.

Zach didn't build in the deep blue at the farthest reaches of the spider chart. He didn't try to solve problems AI can't touch yet. He built just outside the red, in the shallow blue, where the models were good enough, the category was proven, and the timing was right.

Cal AI wasn't science fiction. MyFitnessPal already had a photo calorie feature. What Zach did was execute it faster, with better models, to a younger audience that the incumbents had stopped paying attention to. That's why MyFitnessPal bought it. You don't acquire what you can't build. You acquire what someone else already built, scaled, and owns.

Zach is 19. He didn't have a pedigree, a network, or a decade of experience. He had a point of view and the tools to build on it. That's Creator Capitalism.

Creator Capitalists don’t learn AI. They build with AI. They show up with judgment, perspective, and a problem worth solving—and they use AI to scale what makes them different.

Play with AI in the Far Blue

One person with a point of view and AI can now produce what used to require an entire team. We watched it happen live on our podcast last week when Pirate Eddie uploaded 20 unstructured Creator Capitalist profiles into Claude and produced category science—data visualizations, origin story analysis, career stage breakdowns—in one sitting. Work that would have taken a consulting team a full week.

Eddie didn’t “use AI.”

He was playing with it. Much like an orchestra conductor says, what if we played this much faster and quieter? Or an NBA coach tells Steph Curry he has a green light to shoot from half-court. Or an amazing chef decides, what if I made a savory cannoli?

Pirate Eddie asked Claude (his wife calls it Cleddie), please turn these words into magic numbers. And make it pretty.

He brought a hypothesis. AI brought processing power. Together, they created Intellectual Capital that neither could have produced alone.

If you’re an alumni of the Creator Capitalist course, you can check out the full report in the Mighty Network space.

Don’t Wait for Permission

Some people aren’t waiting for permission.

Mary Johnson—CXO, Mischief Maker, and a Category Design Academy Student—posted this last week:

“I know for a fact that the only reason I haven’t started and exited multiple successful companies is because of my own mental model limitations based on all the crap I have told myself over the years. The person I needed to become who is capable of those multiple exits didn’t exist until now. And she currently knows that NOW is the time for her to emerge and GSD.”

Mary didn’t post about learning a new tool. She posted about becoming a different person. That’s the shift. The obstacle was never the market, the timing, or the technology. It was the story she was telling herself about what she was allowed to become.

She also said something that should make every professional reading this sit up:

“If you aren’t creating net new value by becoming a Creator Capitalist, you will be left behind over the next 24 months and be on the wrong side of the door to this party.”

Twenty-four months. That’s Mary’s timeline. Given what we just saw in those two viral posts—AI theoretically capable of most knowledge work, and a live demo of it replacing a day’s work in four minutes—twenty-four months might be generous.

Name what you see that others don’t

Here’s what we’ve noticed working with hundreds of professionals over the last few years: the people who successfully made the Creator Capitalist shift didn’t start with a business plan. They didn’t start with a website. They didn’t start with a side hustle.

They started with a conversation.

They sat down—with a mentor, a thinking partner, a co-founder, or increasingly with AI—and they answered one question honestly:

What do I see that other people in my industry don’t?

That question is the seed of everything. Your Superpower. Your Intellectual Capital. Your point of view. The problem you were born to solve. Every Creator Capitalist in our book traces their origin back to a moment where they named something nobody else was naming.

Thomas Parrott was a DTC executive who thought his niche was taking US companies to Europe. One jam session later, “Contribution Margin Growth Architect” crystallized—and his sales cycle collapsed from 30 weeks to 30 minutes.

Danny Bauer was a school principal talking about leadership the right way for ten years—but using the wrong words. One throwaway line in the Academy became “Ruckus Makers,” and it acted like a tractor beam for the exact leaders he wanted to serve.

Melissa Andrews and her husband were building an agricultural connectivity company, but the market didn’t know they existed. Then Tom, Melissa’s husband, coined “Digital Darkness” on a plane—two words that captured an entire category. It worked in Australia, New Zealand, the UK, Canada, and the US.

The naming didn’t create their superpower. It unlocked it. Clarity created confidence. Confidence created pricing power. Pricing power created freedom.

Once you have articulated what you see that others don’t, start building with it in the near blue.

Then start playing with it in the far blue.

Build and Play with the Pirate Eddie Bot

The Pirate Eddie Bot is the only AI on the planet trained on the Creator Capitalist framework, the Four Creator Capitals, Category Design, and 200+ Category Pirates mini-books. ChatGPT doesn’t know what a Superpower is in the context of category design. Claude doesn’t know the difference between a Knowledge Worker and a Creator Capitalist. No generic AI can walk you through the specific frameworks in our book—because those frameworks don’t exist anywhere else.

The Pirate Eddie Bot does.

Tell it what you do. Tell it what you’ve built inside your company. Let it show you what you own that you don’t realize you own—and what it could be worth if you codified it under your name. Let it push back. Let it ask you the questions your boss would never ask because the answers would make you too expensive to keep.

The bot won’t just help you see your Intellectual Capital. It’ll help you figure out how to package it, price it, and turn it into something that pays you—whether you make the leap next month or next year.

👉 Founding Subscribers can start jamming with the Pirate Eddie Bot here.

The book drops in six days. But the people who walk into March 17 having already audited their Intellectual Capital—having already seen the number, felt the sting, and started mapping what it’s worth—they’re going to read the book like a blueprint, not a concept.

Start today. The bot is waiting. And it won’t judge your puke.

Creator Capitalist launches March 17.

Join the waitlist at creatorcapitalist.ai.

You’ll get the Introduction and Section 1 opener immediately, early ordering on March 16, and a launch-week bonus we’re announcing soon.

The Anthropic chart showed you the gap. The GPT-5.4 demo showed you the speed. Mary showed you the mindset. The Pirate Eddie Bot gives you the first step.

Six days.

Arrrrrrr,

Category Pirates 🏴‍☠️

Eddie Yoon

Christopher Lochhead