Origami Robotics

Robots today struggle with general-purpose dexterous manipulation, the kind of fine motor skill and manual dexterity humans take for granted. The company's blog post on 'The Dexterity Deadlock' points to a specific technical cause: high-ratio gearboxes, which are commonly used in robot actuators, break sim-to-real transfer, destroy force transparency, and wear out faster than other components. This suggests that existing hardware approaches to robotic manipulation have structural limitations that prevent robots from reliably transferring skills learned in simulation to real-world physical tasks.

Origami Robotics is developing technology aimed at bringing human-level dexterity to robots through general-purpose manipulation systems. Their work involves training manipulation policies and deploying them on physical robots, with an emphasis on building systems that convert robot rollouts into traceable training data and training runs into reproducible, measurable physical evaluations. The company is also addressing hardware-level issues, as referenced in their technical writing about gearbox design tradeoffs affecting sim-to-real transfer and force transparency. Open engineering roles indicate the team works across robot learning, systems integration, control, sensing, and evaluation tooling.