iRobot Founder Wants to Put a Robotic Familiar Into Your Home

2026-05-04 · Source: IEEE Spectrum · Field: Technology & Digital — Robotics & Autonomous Systems, Artificial Intelligence & Machine Learning · Depth: Intermediate, long

Summary

Familiar Machines & Magic (FM&M), a new "physical AI" company founded by iRobot co-founder Colin Angle, has announced its first robot, "Familiar." This quadrupedal social robot, designed for adults and families, features 23 degrees of freedom for lifelike movement and expressive behaviors. It is covered with a custom touch-sensitive coat, a vision system, and a microphone array, powered by an onboard edge AI stack with a small multimodal model optimized for social reasoning. The robot, which resembles an abstracted bear, aims to reinforce healthy routines by perceiving user activities, such as excessive screen time, and actively engaging them in alternative activities. FM&M emphasizes creating long-term value through a "do no harm" philosophy, ensuring positive emotional connections without fostering unhealthy dependency, and keeping all AI processing and data onboard the device without cloud streaming.

Key takeaway

For AI Product Managers developing consumer robotics, you should prioritize designing for long-term engagement and clear value proposition beyond initial novelty. Focus on specific, beneficial routines and ensure robust privacy measures with onboard processing to build trust and avoid the pitfalls that led to previous social robot failures. Your marketing strategy must manage expectations carefully, highlighting what the robot "will do" rather than just its existence as a robot.

Key insights

Familiar is a social robot designed to foster healthy routines and positive human interaction through embodied AI and non-verbal communication.

Principles

• Avoid human-like speech for social robots.
• Design for forgiveness and managed expectations.
• Prioritize onboard AI for privacy and autonomy.

Method

The robot uses a small multimodal model trained on stories, processing speech and vision inputs to generate high-level decisions. These decisions drive a behavior engine and a reinforcement learning unified motion model for socially responsive, non-deterministic behaviors.

In practice

• Use non-verbal cues (tail, ears, eyes) for communication.
• Integrate soft materials and low center of gravity for safety.
• Train AI on stories for plausible intelligence.

Topics

• Familiar Machines & Magic
• Social Robotics
• Human-Robot Interaction
• Edge AI
• Multimodal AI

Best for: Investor, Entrepreneur, Robotics Engineer, AI Product Manager, Director of AI/ML

Related on AIssential

• What Amazon’s Astro Taught Me About Giving Robots a Soul — Character design for embodied AI must be intentional, integrated early, and adapt to context to create truly "aware" products.
• Emma the joke-telling robot cracks up the care home: Paula Hornickel’s best photograph — Social robots offer companionship and engagement in care settings, but cannot replace human connection.
• Do you want to build a robot snowman? — Robotics deployment requires balancing impressive engineering with complex social and brand implications.
• I hate my AI pet with every fiber of my being — AI companion robots struggle to deliver genuine emotional connection despite advanced claims and high price points.
• House Robots Are Coming—and They Will Be Dangerously Cute — Adorable, AI-powered house robots are emerging, designed to form intimate bonds while logging owner information.
• For Robotaxis, Safety Must Be Built In, Not Bolted On — Public perception of robotaxis is highly positive, driven by convenience, safety, and privacy benefits.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by IEEE Spectrum.