You want a sword. It's 1500. You walk into a town and find the blacksmith. He's spent twenty years mastering his craft. He knows the temperature of the forge by the colour of the flame. He knows which hammer strikes shape the blade and which ones shatter it. You're not paying for a sword. You're paying for his decades. His hands. His knowledge that you could never acquire yourself.
You need him. There is no alternative. The sword doesn't exist without the blacksmith.
Now hold that thought.
Someone builds a machine
It's 1800. Someone figures out that you can stamp metal into shape with a press. A factory worker with a week of training can produce more swords in a day than the blacksmith could in a month. The swords aren't as beautiful. They don't have the blacksmith's signature. But they work. They cut. They do the job.
The blacksmith's twenty years of expertise? Irrelevant. Not because his skill wasn't real. It was. But because the barrier he represented has been removed. The person who built the factory didn't need to know how to forge a blade. He needed to know what to build and why.
The value moved. From the hands to the head. From execution to vision. From "how" to "what" and "why."
This is happening right now
I've been building software with Claude Code for months. Not as a developer. I've never been a developer. I'm a marketing and product generalist who has spent twelve years learning how businesses work, what customers want, and where opportunities hide.
And I'm shipping products.
Real products. Not toy projects. Not "vibe coded" demos that fall apart when someone clicks a button. Actual working software that people use. I'm building a game studio. I'm building tools. I'm doing things that two years ago would have required hiring a team of engineers and waiting months for delivery.
The blacksmith is dead. The factory is here. And it fits in a terminal window.
The ideas economy
For the last thirty years, we've built an economy that rewards execution above all else. Can you write the code? Can you design the system? Can you implement the architecture? The "how" people commanded the highest salaries, the most respect, the most leverage.
That's inverting. Fast.
When AI can handle the execution, what's left? Taste. Judgement. The ability to look at a market and see what's missing. The ability to understand a customer's problem deeply enough to articulate a solution before the first line of code is written. The ability to connect dots across domains that specialists never overlap.
The expert generalist is the new power player. Not the person who knows everything about one thing. The person who knows enough about many things to see patterns that specialists miss and direct AI to build what needs building.
This isn't hypothetical. This is my Tuesday.
The resistance
I hear it constantly. "AI can't do what I do." "The code it writes isn't production quality." "You need real engineers for real software." "This is just a toy."
These are the blacksmiths talking.
And they're right. For now. AI-generated code isn't always as elegant as hand-crafted code. AI doesn't understand the nuances of every system. There are edge cases it misses and architectural decisions it gets wrong.
But "for now" is the most dangerous phrase in technology. "For now" is getting shorter every quarter. The gap between what AI can do and what a specialist can do is narrowing at a rate that should terrify anyone whose entire identity is built on a single technical skill.
The blacksmith also made better swords than the factory. That didn't save him.
The permission myth
Here's what really changed. It's not just that AI can write code. It's that the entire permission structure around building things has collapsed.
You used to need permission. Permission from investors to fund a team. Permission from engineers to build your idea. Permission from designers to make it look right. Permission from DevOps to deploy it. Every step required someone else to say yes.
That gatekeeping is gone. A person with a clear idea, good taste, and the ability to articulate what they want can go from concept to shipped product in days. Not months. Days.
The question is no longer "can you build it?" The question is "should you build it?" And that's a fundamentally different skill. That's product thinking. That's market understanding. That's taste.
What this means for you
Stop optimising for depth. Start optimising for breadth. The person who understands product, marketing, design, and technology at a conversational level will outperform the person who has spent a decade mastering one of those disciplines. Not because depth doesn't matter. But because the returns on depth are collapsing while the returns on breadth are exploding.
Develop taste. This is the hardest one to teach and the most valuable skill in the AI era. Taste is knowing when something feels right. It's the difference between a product that works and a product people love. AI can generate infinite options. Knowing which one is right is entirely human.
Learn to direct, not just do. The skill of the next decade is articulation. Can you describe what you want precisely enough that an AI can build it? Can you evaluate the output and know what's wrong? Can you iterate quickly? This is a new discipline and most people haven't started developing it.
Move fast. Speed is the ultimate advantage now. When anyone can build anything, the person who ships first wins. Not the person who builds the best v1. The person who gets something real into the world while everyone else is still writing specs and scheduling sprint planning.
The blacksmith's legacy
I have respect for the blacksmith. Genuine respect. His craft was real. His skill was earned through years of painful, dedicated work. The world owes a lot to people like him.
But respect doesn't change economics. The factory came anyway. The blacksmith who survived was the one who stopped defining himself by his hammer and started thinking about what people actually needed. Some became factory owners. Some became designers. Some became something entirely new.
The same choice is in front of every specialist right now. You can cling to the craft. You can insist that what you do can't be replicated. You can dismiss the tools as toys.
Or you can pick up the new tools, combine them with everything you already know, and build something the pure specialists never could.
The blacksmith is dead. Long live the builder.