Max Cao

2024 is the year of robotics. AI is set to revolutionize the field, leveraging principles from technologies like ChatGPT to achieve remarkable advancements. Early research suggests that Generative AI (GenAI) and foundation models offer the potential for robots that can reason, adapt, and interact in dynamic environments, enabling them to understand complex tasks and operate autonomously with minimal data.

The primary hurdle to realizing this potential is the need for vast, diverse real-world datasets. Unlike ChatGPT, which was trained on centuries of accumulated text, robotics lacks extensive data sources, and current data collection methods, such as paid teleoperation, are unsustainable. While ideas like using simulation data are valuable, the holy grail is large-scale production deployment of robots in diverse environments and use cases.

Moreover, the inherently non-deterministic nature of AI systems raises concerns about safety and reliability. Failure cases for robots, such as collisions with the environment or people, carry great consequences. Ubiquitous adoption is dependent on implementing guardrails to address this.

Jacobi is building software that makes robot deployments scalable and flexible while ensuring the safety of AI-powered robots. Enabled by recent breakthroughs in motion planning technology, Jacobi’s software is applicable to a wide range of use cases and brings a modern software approach to robotics.