Jonathan Curl, a robotics technician at ROBOTIS, discusses physical AI, robot manipulation, industry standards, education, and future applications with the Disney Parks LIVE host.
Curl describes ROBOTIS’s research with MIT on improving robotic grasping. Robots need to sense grip, friction, and how an object might slip, so they can handle objects with different shapes and surfaces. He says physical demonstrations could be combined with AI-generated video to expand training data beyond what robots can practice directly.
The MIT project aims to develop a capable robotic hand that other robotics teams can build on. Curl says it may take 10 to 15 years for some research advances to become reliable at scale, and notes that even basic choices such as the number and placement of arm joints remain unsettled. ROBOTIS’s OpenManipulator line offers reference platforms in different sizes, while companies including NVIDIA are working on hardware compatibility.
Curl says automated public transit could offer more frequent service through driverless trains or on-demand shuttles. The host distinguishes fixed instructions, such as following a route at set speeds, from physical AI that is given a destination and must work out how to get there. Curl also points to hazardous environments as an early robotics application, with disaster response potentially becoming more accessible as systems become cheaper.
Curl compares responsible progress to skateboarding: people should not go faster than they know how to stop. He says companies need to identify unsafe research boundaries. For beginners, he describes ROBOTIS Ollo educational kits, programming options including Blockly-style tools and Python, and a Coach 1V1 robot arm kit developed with Hugging Face for at-home physical AI learning.
ROBOTIS’s Ollo kits are designed for classroom and after-school teamwork, followed by simpler STEAM Cup robots, small humanoid kits, and more advanced Engineer kits. Curl says the products build skills progressively, from block-based programming to Python and industrial actuators. He also recommends open robotics courses on GitHub and university websites for people learning independently.
Briefs summarise talks and interviews; timestamps open the video at that moment. Claims in a talk are the speaker's own.
Brandon, technical specialist and STEM consultant at ROBOTIS, joins Alison Mishra and Todd Borland at AWS to explain motor and servo choices, smart actuator feedback, and controlling a robot arm with cloud messages.
Dr. Dennis Hong of Virginia Tech presents the DARwIn-OP humanoid robot, its development and research goals, alongside collaborators at Purdue, the University of Pennsylvania, and ROBOTIS.