The Robot Report Podcast · 4 May 2026 · 54m
Colin Angle, founder of iRobot, introduces his new venture, Familiar Machines and Magic, which is emerging from stealth to build an empathetic, four-legged robotic companion. The company combines iRobot’s price-performance engineering, Boston Dynamics’ mobility, and Disney’s animatronics to create a pet-like robot with 23 degrees of freedom and a soft, digitally knitted exterior. Angle argues that the future of physical AI lies in emotional connection rather than just industrial efficiency, aiming to ship the first product next year at a cost comparable to owning a pet.
Where this conversation describes AI being used to build or deliver a product. Part of the AI in practice tracker.
Engineering & simulation
The team used AI to expand a dozen initial personality stories into 20,000 vignettes, which were then used to train the multimodal model for the robot's social behavior.
“We start by writing stories about what a familiar, how familiar should act and then use AI to take those dozen stories and turn them into 20,000 vineyards, train the models on the vineyards.”
Speaker 3 · Hear it at 16:08
Engineering & simulationreinforcement learning
The CTO, who previously managed reinforcement learning for iRobot, applied this technology to the new robot's behavior engine to enable autonomous social interaction.
“He was responsible for Gen AI and Marine Force got learning at IRobot. So he put 3500000 reinforcement learning robots, eventually learning on them to improve, uh, room to performance.”
Speaker 3 · Hear it at 33:00
founders before launch
Other shows on the same ground, closest first.
Andrew Kiguel, CEO and Chairman of Realbotix, discusses the company's focus on social companion robots rather than industrial automation, leveraging his background in finance and previous ventures like Hut 8. He details the development of the Aria educational robot, which uses a closed-loop, STEM-based AI to provide safe tutoring in schools, and introduces the new Echo line, including characters like Amar and Mila, designed for stationary, human-facing interactions. Kiguel argues that stationary designs offer practical advantages in battery life and safety, while highlighting the market validation provided by competitors like UBTech and Unitree in China.
Aaed Musa, co-founder of the robotics startup Mecha and a popular YouTube creator, discusses his journey from 3D-printing servo robots in his parents' garage to designing advanced quadrupedal robots. He details the mechanical evolution from high-backlash servos to custom capstan drives using Dyneema DM20 rope, and explains how his startup uses NVIDIA Isaac Sim and reinforcement learning to train expressive, home-friendly robot dogs. Musa also shares insights on his YouTube storytelling process, the importance of industrial design in making robots approachable, and his vision for a platform-based home robot that bridges the gap between toys and industrial tools.
Kinetic artist Sarah Petkus discusses her work building robotic art installations, including the 'Robot Army' of miniature Delta robots and 'Noodlefeet,' a quadruped robot designed to interact with humans. She explains her transition from illustration to mechatronics at SAIC, detailing how she used Algodoo for simulation and 3D printing for rapid iteration. Petkus also covers the business side of her hobby-turned-company, Robot Army LLC, including their Kickstarter campaign and the challenges of self-promotion as an introvert.
Hosts Daniel and Forbode discuss the M4, a bio-inspired multimodal robot developed by Caltech and NASA JPL that can fly, roll, walk, and tumble using eight different modes of locomotion. They highlight how the robot's AI autonomously selects the best traversal method based on terrain, drawing inspiration from nature such as birds and sea lions. The episode also covers Jonathan Schultz of Team Huge, whose BattleBots design philosophy of accessible core components parallels the M4's adaptive engineering, while exploring applications in space exploration, emergency rescue, and last-mile delivery.
Piotr Owczarek, President of Fideltronik, discusses the company's seven-year journey to develop AI-driven robotics for electronic manufacturing services (EMS). He explains how a team of 170 engineers in Krakow is creating flexible automation that mimics human operators to address labor shortages and cost pressures. Owczarek details their 'academy for robots' concept, where collective knowledge is stored in the cloud to allow robots to learn from each other without daily programming. The episode highlights that major EMS companies are currently engaged in pilot programs to validate this technology before full-scale adoption.
The summary, chapters and topics are generated from the episode’s audio by this directory, not written by The Robot Report Podcast. The episode itself belongs to its publisher.