Most people think AI is a step change in manufacturing. It’s not. It’s the third wave of a 150-year progression that can be broken into 3 waves:
Each wave didn’t just improve efficiency, it changed what was even POSSIBLE to build. We’re now living through Wave 3.
Here’s the framework:
1️⃣ Wave 1: Mechanical Automation (late 1800s → early 1900s)
This was about moving things. Conveyors. Assembly lines. Electric motors.
We mechanized physical motion at scale (lifting, transporting, assembling).
→ Output exploded
→ Labor became specialized
→ Manufacturing scaled for the first time
But systems were rigid. They could only do one thing, one way.
2️⃣ Wave 2: Computer-Controlled Automation (1970s → 2000s)
This was about precision and programmability. CNC machines. Industrial robots. PLCs.
We moved from mechanical systems to software-defined control, unlocking complex, repeatable tasks.
→ Skilled work (machining, welding, inspection) became automatable
→ Factories became programmable (sort of...we're still working on that!)
→ Flexibility increased
But these systems were still deterministic...they only did what they were explicitly coded to do. No adaptation. No judgment.
3️⃣ Wave 3: AI-Driven Automation (2010s → now)
This is about decision-making.
For the first time, machines can:
→ Perceive (vision, sensors, messy data)
→ Adapt (variation in materials, environments)
→ Decide (optimize in real time)
This pushes automation beyond the shop floor into: procurement, logistics, quality, design. customer communication.
All the messy, unstructured work that used to require humans.
The key insight:
Wave 1 automated motion
Wave 2 automated precision
Wave 3 automates judgment
That’s why this moment feels different. Not because automation is new, but because we’re crossing a boundary we’ve never crossed before.
We're heavily invested in this wave at Baukunst.
Our latest study group report from the collective dives much deeper into this. I'll drop it in the comments!
