All Recordings

VLA vs WAM: Do Robots Need to Predict the World, or Just the Next Action?

About This Talk

Debate whether robots need world models or just the next action. Panelists from Agility Robotics, Walden Robotics, Shield AI, NVIDIA, and Bonsai Robotics compare VLAs and WAMs for real-world autonomy. From Actuate, Foxglove's annual robotics developer conference.

Speakers

Chris Paxton, AI Innovation Lead, Agility Robotics

Chris Paxton is an AI + Robotics Research Scientist at Agility Robotics, where he works on robot motion and task planning, integrating algorithms with perception and learning to build intelligent, adaptive robotic assistants. He previously held research roles at Meta AI (FAIR), NVIDIA, and Zoox, and holds a PhD in Computer Science from Johns Hopkins University. Chris is also the co-host of the RoboPapers podcast and writes the robotics newsletter "It Can Think" on Substack.

Ben Burchfiel, Co-Founder & CTO, Walden Robotics

Ben Burchfiel is CTO and co-founder of Walden Robotics, the full-stack Physical AI company that spun out of Toyota Research Institute and launched from stealth in July 2026 with roughly $300 million in seed funding at a $1.1 billion valuation, with its general-purpose robots already working production shifts at a Toyota plant in North America. He previously led the Large Behavior Models division's R&D at Toyota Research Institute, where he headed the multimodal policy learning and data teams building general-purpose robots via AI foundation models trained at scale — foundational work, including Diffusion Policy and LBMs, that now powers Walden's robots.

Armon Shariati, Staff Software Engineer, Shield AI

Armon Shariati is a Staff Machine Learning Engineer at Shield AI, where he develops advanced autonomy that enables intelligent systems to operate in complex real-world environments. Before joining Shield AI, he served as an Assistant Professor of Robotics and Controls Engineering at the U.S. Naval Academy and as an Applied Scientist at Amazon Robotics, where he developed machine learning algorithms for scalable decision making and robust real-world autonomy. Armon earned a Ph.D. in Computer Science and a Master of Science in Computer and Information Science from the University of Pennsylvania, as well as a Bachelor of Science in Computer Engineering from Lehigh University.

Danfei Xu, Research Lead, NVIDIA

Danfei Xu leads research on robot learning from human data at NVIDIA GEAR. He is also a faculty member at Georgia Tech, where his research broadly focuses on robot learning and planning. His work has received Best Paper Awards at CoRL, IEEE Robotics and Automation Letters, and ICRA. He received the NSF CAREER Award in 2025.

John Macdonald, Head of AI, Bonsai Robotics

John Macdonald is the Head of AI at Bonsai Robotics. Convinced of the potential of robotics from a young age, he started in competitive robotics in grade school and hasn't looked back. Previously, he worked on deep learning-based obstacle avoidance and 3D reconstruction at Skydio. Before that, he worked on autonomous organic crop navigation & monitoring at Carnegie Mellon University. John holds an M.S. in Robotic Systems Development from Carnegie Mellon University and a B.S. in Computer Science from Cornell University.

More Recordings