Dimensional Hires Humans to Walk Robot Dogs in City
The startup is collecting navigation data for its dimOS software as current AI models fail to pilot hardware autonomously.
Updated on Oct. 28, 2026 in Robotics
Live Poll
Do you believe robots will successfully replace most manual labor jobs in your lifetime?
On October 5, 2026, the San Francisco startup Dimensional began hiring workers for $20 per hour to manually operate and walk its fleet of 100 robot dogs. The company is currently using human labor to gather environmental data while its artificial intelligence agents struggle to navigate urban spaces autonomously.
Why it matters
The reliance on manual labor highlights the current limitations in AI-driven robot navigation, as firms pivot from purely digital training to real-world data collection. Dimensional aims to use this human-generated data to refine its dimOS operating system for future deployments in factories and data centers.
Dimensional purchased 100 robot dogs from Unitree at a cost of roughly $2,000 per unit. Each 33-pound unit requires human operation via mobile app or remote to record sensor data during four-to-six-hour shifts.
The players
Dimensional
A San Francisco startup based in the Dogpatch neighborhood developing dimOS, an open-source operating system for industrial and utility robotics.
Unitree
A robotics manufacturer that supplies the hardware platforms for Dimensional, including the $2,000 quadruped robots currently undergoing data training.
The details
Workers guide the 33-pound robot dogs using a mobile application that logs environmental sensor telemetry. This data serves as the foundation for training AI agents to process natural-language instructions for physical tasks. The company's open-source operating system, dimOS, is designed to eventually shift these robots toward autonomous navigation in complex environments like construction sites.
Timeline
October 5, 2026: Dimensional posted the job advertisement for robot walkers.
October 28, 2026: Current date of the report.
The Tech Race
Dimensional is currently attempting to overcome the limitations of current-generation AI navigation by using human-operated data collection. This approach reflects a broader industry push to move beyond lab environments to train AI for industrial, construction, and data center roles.
Residents in the Dogpatch neighborhood may see staff conducting four-to-six-hour shifts of robot operation in public or testing areas. The project remains in the data-gathering phase, with no current timeline for when these units will transition to autonomous operation.
The takeaway
Dimensional is currently prioritizing human-guided data collection to bridge the gap in autonomous navigation for its robot fleet. Observers should monitor the release of version updates for dimOS as a benchmark for when the startup achieves consistent, non-supervised robot performance.
Further reading
For more on the development of physical hardware and control software, visit the Robotics section.
Live Poll
Do you believe robots will successfully replace most manual labor jobs in your lifetime?
