Wayve aims to commercialize its system at an ADAS level first, similar to Tesla's approach, by leveraging a widespread rollout to collect data for achieving full autonomy. Unlike Tesla, Wayve is open to incorporating lidar for near-term full autonomy. Kendall introduced GAIA-2, a generative world model that enhances the AI driver's adaptability and human-like driving behavior by processing video, text, and actions together. Wayve shares a philosophy with autonomous trucking startup Waabi, focusing on scaling data-driven AI models that generalize across different environments and using generative AI simulators for testing and training.
Key takeaways:
- Wayve's strategy focuses on creating autonomous driving software that is cheap to run, hardware agnostic, and applicable to various systems, including ADAS and robotaxis.
- The company plans to license its self-driving software to automotive and fleet partners, emphasizing its ability to work with existing sensors and GPUs in vehicles.
- Wayve's approach to autonomy is similar to Tesla's, using an end-to-end deep learning model, but it is open to incorporating lidar for full autonomy.
- Wayve's GAIA-2 model uses generative AI to train its driver on real-world and synthetic data, aiming for adaptive and human-like driving behavior without hand-coded instructions.