We’ve mistaken where progress really comes from. It’s not the invention—it’s the reinvention of work.
We’re pouring billions into AI, automation, and other “hero” technologies, hoping for a productivity miracle. But the real source of past leaps wasn’t the tech itself. It was how we reorganised work around it. In my latest Substack post, The Power Loom Principle, I explore how this blind spot is stalling growth—and what we must do to reignite it.
Productivity growth has flatlined. Firms are desperate, policymakers are anxious, and the usual response—invest more in breakthrough technologies—isn’t working. The assumption is clear: innovation flows from science to tech to productivity. If we just build the right tool, the gains will come.
But what if we have it backwards?
The technologies that defined past revolutions—electricity, combustion, even the power loom—didn’t deliver their full value on arrival. They became transformative only when we restructured how people worked around them. The productivity boom of 1870–1970 wasn’t just a lucky run of invention. It was a shift in work systems.
The power loom wasn’t remarkable for its technical brilliance. Its true impact came when it forced a reorganisation of labor—from home-based weaving to the factory floor. That change enabled each worker to oversee multiple looms, instantly multiplying output. It also redefined who held expertise, embedding knowledge into tools and systems, not just people.
This is the Power Loom Principle: Productivity surges when we redesign the system of work, not just the tools we use. It’s a principle that explains why certain technologies trigger explosive growth while others fizzle. And it’s exactly what we’re missing today.
We’ve fallen for a flawed productivity logic:
- Productivity comes from innovation
- Innovation comes from technology
- Technology comes from research
That neat chain makes us overinvest in R&D and underinvest in the messy, ground-level transformation of work. Worse, it leads us to treat “hero” technologies—like LLMs or autonomous vehicles—as silver bullets, expecting them to save us without changing how we operate.
But history tells a different story. Technologies become productive after we reorganise work around them, not before. Thermodynamics came after the steam engine. Science explains what practice has already invented. Learning by doing—not research papers—is where the real breakthroughs begin.
In sector after sector, real gains come not from theory, but from practice:
- Pilots and speedbugs: A small change in cockpit instruments externalised memory, freeing pilots to focus on flight and embedding expertise in the system.
- Surgeons and checklists: Codifying routines reduced reliance on heroic recall and made hospitals safer.
- Coders and GitHub Copilot: By redistributing coding knowledge into AI suggestions, junior developers can contribute faster, and teams scale expertise.
Each case represents a shift: not in the tool alone, but in how knowledge is embedded, shared, and used. The system—not the genius—is the innovation.
Take autonomous taxis. Hailed as the next great productivity leap, they’re the modern equivalent of the power loom. But like the loom, their impact won’t come from the tech itself. It will come when we redesign how work is done:
- A single remote operator overseeing 10–50 cars.
- New interfaces, oversight protocols, and task structures.
- Reallocation of labor across space, not just automation of it.
Until we do that, AVs will remain expensive novelties. But when we move the boundary of the work system—just like when weaving left the home for the factory—we’ll unlock a multiplier effect.
Economist Robert Gordon saw the 1870–1970 boom as a one-off, driven by a lucky run of “general purpose technologies.” But he missed the point. These were not one-time technologies—they were one-time reorganisations of work systems. Once you’ve moved the boundary of how work is done, you can’t move it in the same way again.
Today, we keep searching for the next hero tech. What we need is the next boundary shift. Until we transform our work systems—how knowledge flows, how tasks are structured, how interfaces are designed—no amount of AI will save us.
We’re not suffering from a shortage of innovation. We’re suffering from a failure of imagination—about how work itself must change.
The real challenge isn’t to invent new tools. It’s to reimagine the systems we embed them in. That’s how we got the factory. That’s how we’ll get the future.
I explore this in depth in The Power Loom Principle.