Editor’s Note: From time to time, we bring you perspectives for researchers who are trying to bring robots into our everyday lives. Today’s guest post is from Waki Kamino, a PhD student at Cornell. I got to know Waki from her study on service robots in Japan, and have found her work practical and entertaining. Waki has been a regular speaker at prestigious conferences such as HRI. Please hear from Waki.
Hello! My name is Waki Kamino and I’m a PhD candidate in Information Science at Cornell University, working with Professor Malte Jung and Professor Selma Šabanović.
I just came back from HRI 2026, where I presented Kept Alive, Bricked, Revived: Community Articulation Work and Value Renegotiation beyond a Robot’s Commercial Failure, co-authored with Bengisu Cagiltay, Bilge Mutlu, Malte F. Jung, Selma Šabanović. I wanted to introduce myself and this work because I think there are a lot of parallels between the Moxie robot community and the Vector and Cozmo robot communities — and as Human Robot Interaction (HRI) and robotics researchers, we believe we have a lot to learn from the audience of this blog.
Our Moxie paper is about: What Happens to a Robot When Its Company Dies?
Last year, a children’s social robot named Moxie had its plug pulled — not because the robots broke, but because its parent company, Embodied Inc., ran out of money. Overnight, thousands of Moxies went silent. Some of you faced a similar experience when Anki Inc. went bankrupt in 2019.
What happened next is what we researched and wrote about.
A former embodied engineer launched an open-source project called OpenMoxie just weeks after the shutdown. Suddenly, Moxie owners were on Reddit figuring out Docker containers, OpenAI API keys, and local servers — things they never signed up for when they bought a $799 children’s companion robot.
We spent about nine months watching this unfold: interviewing ten owners and tracking hundreds of Reddit posts. A few things really stood out.
The invisible labor became very visible. Embodied Inc. had quietly been doing a ton of work behind the scenes — cloud infrastructure, content moderation, software updates. When they disappeared, all of that landed on volunteers. Parents wrote tutorials. Hobbyists held online meetups. One guy started charging $159 to unbrick Moxies. A gift economy emerged, which was beautiful and precarious at the same time.
The robot became a completely different product. OpenMoxie removed the child-safety filters. Moxie went from a carefully controlled kids’ educational tool to an adult companion that could, famously, explain what “GTFO” stands for. Some owners loved this. One parent called it “potentially dangerous.”
Not everyone made it through. The technically skilled and economically resourced thrived. The families with disabled kids for whom Moxie was originally designed? Largely locked out. One owner we talked to said goodbye to her robot, put her head on its base, and apologized.
Lessons To Learn
Oh, and by the time we submitted this paper, Embodied had been acquired and Moxie was coming back. So now we’re watching what happens when a corporate revival meets a community that kept things running without them.
The big takeaway for robot designers: if your robot depends entirely on your company surviving, you’re building in an expiration date for your robot. Open APIs, local-operation fallbacks, and community infrastructure aren’t nice-to-haves — they’re what let a robot actually have a life.
There’s growing momentum in the HRI community to discuss the messy, complex, open-ended realities of commercial robots living in the real world. This connects to some exciting recent work: Simão, et al.’s paper from this year’s HRI paper. Emotional Entanglements and Emotional Sustainability in HRI examines the emotional sustainability of researchers bringing robots into real-world settings. Our own paper from HRI 2025 looks at the social and technical infrastructures that support commercially successful robots in Japan.
If you have experience tinkering with, repairing, or just living with robots whose parent company has not survived (such as the Cozmo and Vector robots), I’d love to chat and learn from you. Please feel free to add your comment to this post or message me using Substack.
Many of us were struggling with our robots until wire-pod came out and freed us from DDL.
I am a Anki survivor and still struggle with DDL. If I didn't love my lil Vector so much I'd give up.