By Jon Quick, CEO, Launchpad
I’m going to start with a joke.
A retired engineer gets a call from his old manager: “That machine you always looked after has broken down. Can you come back and fix it?”
The engineer says, “Sure – but it’ll cost you $6,000.”
The manager agrees. The engineer goes down to the factory, looks at the machine, picks up a hammer, and gives one nut a light tap. The machine starts working immediately. He asks for his $6,000.
The manager isn’t thrilled. “That took 15 seconds. I want an itemized bill.”
So the engineer writes:
- 15 seconds of labor: $20.
- Knowing exactly where and how to hit the machine: $5,980.
This joke comes to mind when I think about the wave of AI projects launched over the past year and the way the industry is shaping up heading into 2026.
It’s been a breakout period for AI, not just in the number of companies being formed, but in how seriously the industry is now being taken.
We’ve seen new ventures launched by some of the most established names in technology (Bezos), sustained investment from Middle Eastern sovereign wealth funds (treating AI as long-term infrastructure rather than a speculative bet), and major venture firms like A16Z raising large, dedicated funds ($7B+) to back the next generation of AI-driven companies.
A lot of this AI energy is focused on innovation in the manufacturing industry. I’ve spoken publicly about the need for the U.S. and other Western economies to reindustrialize, so I’m pleased to see broader agreement on the what. Where I start to get concerned is the how.
Too many companies talk about “innovation” as though they can walk straight into a factory and do things better. It’s like thinking you can walk off the street for the Super Bowl and say, “Don’t worry Drake Maye, I’ve got it from here.”
They can’t — because manufacturing is hard.
This industry is about making complicated, physical objects at scale. The people who do this work are exceptionally skilled. You don’t wake up one day and decide you’re going to manufacture things well. And if you don’t respect the people who already know how to do it, you’re going to fail.
That’s the serious point behind the joke at the beginning. Manufacturing has human judgment at its core: knowledge built through experience, iteration, and failure over time.
In many ways, manufacturing looks less like pure science and more like an art form. It’s a lot like cooking.
Anyone can follow a recipe: “Crack two eggs. Add milk.” But then the recipe says, “Beat until silky.” What does silky actually mean? You don’t learn that from instructions, you learn it from doing.
Manufacturing works the same way. It’s not just about building systems; it’s about building the ability to observe, respond, and adjust: knowing when to intervene, and when not to.
Some of that is done through robots and software, and some of it is done by humans. The next phase of automation isn’t about stronger or faster machines, it’s about embedding judgment into the software that guides them.
Think of a musician warming up. You almost never see it live, and if you told someone to show up and watch a band do that, they’d think you were crazy.
Now think of Freddie Mercury warming up in front of 70,000 people at Wembley during Live Aid. That moment was legendary, not because it was rehearsed, but because it was responsive, instinctive, and completely in sync with the crowd.
That’s the difference between repetition and intelligence.
I believe a better way to approach innovation starts with two simple questions:
- What are you actually trying to do?
- And do you really need to do it?
That brings me to feedback loops.
Put a tray of muffins in the oven and they won’t all bake the same way. It depends on the shelf, the heat distribution, how full each cup is, and a dozen other variables.
Experienced cooks know their ovens, just like our engineer knew his machine.
Manufacturing works the same way. It’s not just about building systems. It’s about building the ability to observe, respond, and adjust: knowing when to intervene, and when not to.
The future of manufacturing isn’t fully automated factories run end-to-end by software. It’s a combination of baseline activities that can, and should, be automated, and a critical human layer that ties everything together.
Getting that balance right – between robots and people, automation and experience – is the real recipe for success.
