I keep seeing robotics framed as a capability race.
Better perception.
Better manipulation.
Faster inference.
But once robots start doing real economic work, intelligence stops being the interesting variable.
Incentives take over.
The moment a machine participates in markets — moving inventory, running inspections, executing logistics — its performance isn’t judged in isolation. It’s judged against cost curves, time pressure, margin targets. And that pressure shapes behavior whether we admit it or not.
Optimization isn’t neutral. It bends toward what gets rewarded.
That’s the part that made me pause with Fabric.
If robots are going to operate inside shared economic systems, the incentive layer can’t stay invisible. Who benefits from higher throughput? Who pays when a corner gets cut? What defines “efficiency” when speed and safety compete?
These aren’t abstract governance debates. They’re embedded in architecture.
Right now, most robotic systems optimize inside silos. Vendors push updates that improve metrics that matter to them. Operators tweak performance to protect margins. Over time, those incentives compound quietly. You don’t notice the drift until something breaks.
Fabric seems to assume that once machines start participating economically, the coordination layer has to make those pressures legible. Identity, settlement, participation — not as add-ons, but as part of the base infrastructure.
That doesn’t solve incentive tension. It surfaces it.
And surfacing it might be the only way to prevent behavior from drifting toward whatever is easiest to reward.
There’s still risk here. Economic layers can centralize. Dominant actors can steer optimization indirectly. “Open participation” can quietly narrow if incentives aren’t balanced carefully.
But ignoring the incentive layer doesn’t make it disappear. It just hides it.
Robots don’t just get smarter.
They optimize for what the system rewards.
The question isn’t whether machines evolve.
It’s whether the economic structure guiding that evolution is visible — or invisible.