February 23, 2026
Mud vs Wi‑Fi: place your bets
Show HN: Sowbot – open-hardware agricultural robot (ROS2, RTK GPS)
Open farm robot sparks cheers, cloud fears, and muddy reality checks
TLDR: An open, two-brain farm robot called Sowbot aims to make sustainable agriculture easier with precise GPS and shared designs. The community cheers the vision but debates cloud reliance and demands muddy, winterfield testing, asking bluntly: where are the seeds and can it survive real farms?
Sowbot, the open farm robot, rolled onto Hacker News promising greener fields and less grunt work—then the comments took the wheel. The crowd loved the branding—“Great name!”—but instantly split into camps: the future-is-here hype squad vs. the have-you-met-mud? realists. One hot thread blasted the hint of a web app dependency, with users begging for local control over anything that touches the tractor. Meanwhile, farm veterans issued a rainy-season reality check: take it to real fields now, because winter soil doesn’t care about your demo video.
Under the hood, Sowbot touts two “brains” (one for movement and safety, one for cameras and AI), centimeter-accurate positioning via RTK GPS (super-precise satellite guidance), and ROS 2 (a robot software toolkit). It’s fully open hardware, with all schematics and code shared—so much so that a commenter even offered a plug-and-play driver for the on-board sensor. But practical questions cut through the buzz: Where are the seeds? How do they drop? What’s the payload?
The vibe? Optimism meets mud-splattered skepticism. Fans say it’s the path to sustainable farming. Critics want proof it can survive rain, bad signal, and the chaos of real fields. The meme of the day: “Can Sowbot find a carrot—and a Wi‑Fi signal?”
Key Points
- •Sowbot’s Open AgBot provides an open-hardware reference design and a developing production-ready software stack for agricultural robotics.
- •The compute unit uses two Avaota A1 SBCs: Board A for control/safety (ROS 2, EKF) and Board B for perception/AI (camera drivers, preprocessing, YOLO).
- •Native CAN bus and dual GNSS RTK receivers enable robust communication and centimeter-level positioning for field operations.
- •All schematics, PCB layouts, and firmware are open-licensed; the system is housed in a rugged, waterproof aluminum enclosure.
- •The platform aims to cut ~18 months of R&D for startups and enables reproducible research via Docker images, with a roadmap available on GitHub.