Last weekend, we traveled to San Francisco for the Y Combinator x Google DeepMind Multimodal Frontier Hackathon. We're pleased to announce that we became finalists in the Top 4 out of 326 builders!
The hackathon was about pushing AI applications past the "chatbot" phase and into real-world systems using multimodal reasoning.
What We Built
We started building an MVP for a problem a lot of robotics engineers face all the time. Currently, AI cannot understand the full context of robotics development across software, electronics, and hardware. What we're building is a foundational AI model that understands all of this. For example, it would be able to understand how a certain software package runs on a specific electronic component with some specific hardware and how all of this fits together in simulation. This AI would be the backbone to our RDE (robotics development environment) to help robotics engineers develop across all modalities and maintain full context.
What's Next
With this first sense of validation, we're building our first product, Forge RDE, where AI can reason over code, datasheets, CAD files, and live robot behavior all in one place. The goal is simple: make building robots feel as natural as building software.
Thank You
Huge thanks to Y Combinator and the Google DeepMind team for putting together one of the most interesting hackathons we've been to.
We're also grateful to everyone we met that weekend who is thinking seriously about the future of multimodal AI.
Now back to building the future of robotics development tools.