Robotics
Machines that sense, decide and act on the physical world, which is a harder problem than software because the world does not hold still.
Reviewed July 16, 2026Stable
When not to use it
- As a synonym for automation. Much automation is fixed machinery with no sensing or decision at all.
- Where the environment cannot be modified and the task varies, which is where the field is weakest.
- For tasks whose difficulty is cognitive rather than physical; the robot is not the constraint there.
Reach for something else instead
- Fixed automation — purpose-built machinery where the task never changes, usually cheaper and more reliable than a programmable robot.
- Teleoperation — a person controls the machine remotely, which sidesteps autonomy entirely and is a legitimate answer for rare or high-stakes tasks.
- -
Read more on the blog
- 621,000 robots installed, virtually no humanoidsTerritory 7 opens on the question article 117 left unresolved. Industrial robotics is enormous and growing. The general-purpose machine that would move the boundary has almost no deployments.
- 220 million miles, inside a boundary Waymo drewThe strongest safety evidence in physical autonomy, and the methodology that makes it honest is also what limits what it can tell you.
- Robots weed ten million acres. They still cannot pick.Agricultural robotics has a clean split between what scales and what does not, and the line falls exactly where the thesis of this territory predicts.
- The robots removed the walking. It was also the rest.The largest robot deployment on earth, and two sets of injury figures pointing in opposite directions. Both may be accurate, and what they share is more interesting than what they dispute.
Further reading
- International Federation of Robotics, World Robotics 2025 — installation counts and operational stock.
- Geirhos et al. (2020), Shortcut Learning in Deep Neural Networks — why learned policies fail on variation the training distribution concealed.
Primary sources, listed so you can check the claims on this page rather than take them on trust.
Where people go wrong
- Reading a demonstration as a deployment. The gap between the two is where most robotics claims fail.
- Assuming industrial success transfers to unstructured settings, when the industrial base succeeds by removing exactly the variation that unstructured settings contain.
- Ignoring the cost of environment engineering, which is often the largest line in a working installation.
At a glance
FieldApplied AI
Three partsperception, planning, control
Installed base4.66 million units, 2024
Dominant applicationmaterial handling
Defining difficultysim-to-real gap
DifficultyIntermediate
Where this sits
A starting point. Nothing needs to come before it.
0Levelstarting point
0Needs firstnothing
0Opens upnothing further
1Areastays here
Computed from the prerequisite graph, not assigned. How this works