May 27, 2026
Fast drones, faster hot takes
Show HN: Open-Source AI Racing Harness
Drone racing fans cheer the free sim, then instantly ask why it’s not using the big-name tools
TLDR: Elodin released a free practice simulator for a high-stakes autonomous drone race so teams can start early instead of waiting for the official software. Commenters were intrigued but split between skepticism about building custom tools and a very practical demand for a simple starter bot people can try immediately.
A shiny new open-source practice rig for the $500,000 AI Grand Prix has landed, and the vibe is equal parts "finally!" and "okay, but why this setup?" Elodin says its new AI Grand Prix harness lets teams start building autopilot software for autonomous drone racing before the official competition simulator arrives. Translation for normal humans: instead of waiting around, people can start teaching virtual drones to fly through race courses now.
But in classic internet fashion, the applause lasted about three seconds before the comments turned into a mini cage match. One of the strongest reactions came from the "why reinvent the wheel?" crowd, with a commenter bluntly asking why they didn’t just use well-known simulation tools like Gazebo or Isaac. That’s the real drama here: some readers see this as a generous head start for competitors, while others instantly smell a tech flex and want to know why the team built its own stack instead of grabbing an off-the-shelf option.
Then came the practical crowd, delivering the most relatable energy in the thread: "Cool, but can I run something simple right now?" One commenter asked for a basic starter bot, like a line-follower, so newcomers can actually see the simulator work without diving into the deep end. It’s a very Hacker News kind of moment: one person questions the philosophy, another asks for the tutorial, and together they accidentally sum up the whole internet reaction — impressed, skeptical, and mildly impatient for a demo.
Key Points
- •Elodin open-sourced a practice rig for Anduril’s AI Grand Prix so teams can begin developing autonomous drone autopilot code before the official Virtual Qualifier 1 simulator is released.
- •The company says the rig is open source, works on macOS and Linux, and can be set up with uv sync and a short Betaflight build.
- •Elodin presents the release as part of a longer effort to build aerospace simulation tooling that can reduce setup time from weeks to hours.
- •Its stack includes a Rust ECS and JIT-compiled physics core called nox, Python bindings, a 3D editor tied to elodin-db over TCP, and the s10 process runner.
- •The AI Grand Prix practice rig includes 6-DOF rigid-body physics, motor dynamics, drag, ground constraints, multi-rate sensors, and a GPU-rendered 640×360 camera tilted +20° to match the VADR-TS-002 specification.