Was last week’s intro too much of a downer? As we hit “send” on the final Automated newsletter of the year, I’m happy to report that 2025 has been a good year for me, both personally and professionally. Thanks to all of you for helping make this publication and its eponymous podcast day one successes, and if the number of last minute opportunities that have sprung up in the past couple of weeks are any indication, 2026 is gonna be a wild one.
I’m off next week. We are, however, dropping one final episode of the podcast, so you can grab a nice cup of cocoa, snuggle up by the fire, and let us gently break it to your extended family that they’re probably not getting fully functioning home humanoid robots in their stockings next year. Like much of Automated, we went into the episode with a loose idea, and things came together better than any of us expected.
It's our first in-person episode, recorded at Toyota Research Institute’s Bay Area offices. Remember those top 10 robot videos we had you vote on a few weeks back? TRI’s Erin McColl and I channel our inner Tom Servo and Crow and discuss why each is emblematic of the past year in its own way.
So, for all the industry "course correction" we discussed last week (and, spoiler: a bit more below), the selections offer a nice bit of insight into where robotics really stands as we enter this new year.
For me, 2025 was notable for the fact that I changed jobs for the first time in nearly a decade. For the topic I cover, this was the year robotics as an industry began to enter mainstream consciousness in a major way. It's been fascinating to play a role in exposing more and more people to the ins and outs of the subject in much the same way I have been over the past decade.
There's tremendous opportunity to get people excited about the potential, and to engage younger generations in ways that extend beyond the sci-fi of our youths. There's space for robots and automation to do good for the world and the creatures that occupy it. There's also a lot of opportunity for misinformation — for setting unrealistic expectations about what these systems can and can't do.
Some of the most rewarding conversations I've had at Automated are ones that recalibrate. It's possible to be a pragmatist and be excited at the same time. I often encounter roboticists who are still wide-eyed and engaged after all these years. I've also been a journalist for a very long time and, well, you mostly can't really say the same for my people, bless our jaded hearts.
If 2025 was the year the door truly opened for robotics, let's kick off 2026 by helping the people around us separate real systems from their sci-fi counterparts. Doing so will help them better appreciate just how much work and history it took to get us here in the first place.
There are a million lessons from every Chapter 11, some more universally applicable than others. All are, of course, easier to litigate in hindsight. The struggles that preceded iRobot’s own announcement run the gamut. The global pandemic and its years-long knock-on effects almost certainly hastened and compounded many, while introducing entirely new challenges for the New England robotics firm.
Any examination of iRobot’s 35-year history must first acknowledge the implausible nature of the company’s existence. “Lightning in a bottle” undersells it. For decades, it’s been held up as the sole exception to conventional wisdom around home robots. The garages of Silicon Valley are lined with the ghosts of exceptionally intelligent foolhardy folk who failed to crack the code.
After years of robot projects that never quite took, iRobot finally stuck the landing with Roomba in 2002. The ragtag operation that no VC wanted to touch for years bested Electrolux, the Scandinavian appliance giant whose contemporary effort at a robot vacuum overshot the consumer market. As Roomba co-inventor Joseph Jones puts it in his recent book, Dancing with Roomba, “It’s unlikely anyone would have deliberately chosen the path we followed.”
Common wisdom holds that somewhere in the neighborhood of 80-90% of startups fail in their first three years. That number jumps to 97% when zeroing in on hardware. As someone who has covered both that broad category and the subset of robots, I have to assume the platform gets even shorter when talking about the latter. By each of these metrics, iRobot is an extended success story. Here we are in 2025, still uncertain of the company’s ultimate fate.
You wouldn’t know it by reading just about any coverage of iRobot over the past week, but Chapter 11 isn’t necessarily the end. It’s not the ideal outcome, or even a good one. It’s bad. It’s bankruptcy, and generally speaking, you would prefer not to be forced into a position where you have to declare it. But it doesn’t mean the buzzer’s buzzed and time is definitively up.
With Meta nearly in the rearview, AI rockstar Yann LeCun this week confirmed his involvement with Advanced Machine Intelligence Labs (AMI). The French researcher shared the news via LinkedIn, following a report from The Financial Times that the AI startup is looking to raise well over half-a-billion USD. LeCun, who was also recently named an advisor for robotics startup UMA, clearly has a full dance card these days and didn’t waste any words with the post, writing, “A partnership between Nabla and Advanced Machine Intelligence (AMI Labs). Yes, AMI Labs is my new startup. I'm the Executive Chairman. And Alex LeBrun is transitioning from CEO of Nabla to CEO of AMI Labs!”
AMI looks to commercialize LeCun’s longstanding interest in “world models” — AI systems whose ability to form causal relationships more closely mirror human learning than the brute force approach of LLMs (large language models). While the notion of these reasoning-based models has been around for decades, LeCun still managed to make waves when he noted in February at an AI summit, “Never mind trying to reproduce human intelligence. We can’t even reproduce cat intelligence or rat intelligence. Any house cat can plan very highly complex actions.”
With 100,000 hours of kitchen experience to draw from, the new system from Chef Robotics has enough data to satiate 10 Malcolm Gladwells. Chef+ sports ingredient containers that are double the size of its predecessor, while still managing to occupy less overall space on the kitchen floor. The upgraded robot has improved CPU and GPU processing, along with a three-camera vision system, for more accurate ingredient tracking. The news comes nine months after Chef raised north of $40 million to increase system scaling and deployment. The San Francisco-based food prep automation firm says it’s been testing these systems with existing customers. Other interested parties can purchase one through the startup’s site now.
No, I will not be attending next year's CES, for the first time in roughly... honestly, I don't know how long. It's a reprieve I'm mostly happy about, save for Hyundai bringing the Electric Atlas to Vegas's big event.
An eventful year ends on a bit of a dark note, as citywide blackouts cause Waymo vehicles to stall in San Francisco streets over the weekend. The self-driving car company says service resumed as of Sunday evening. Looks like we'll be waiting some time for true lights-out autonomy.
According to information from the Information, Figure head figure, Brett Adcock, plunked $100 million to launch a new AI research lab called Hark.
Now Playing on Automated Pod
Ben Procter (Avatar: Fire and Ash)- The production designer discusses how real-world robots influenced the sequel’s autonomous systems and mechs.
For more than 30 years, Professor Holly Yanco has been a leading mind in human-robot interaction and search and rescue robots.
Grace Brown (Andromeda) - For Grace Brown, a humanoid robot future is a hopeful future. Andromeda’s Abi is designed to build connections in an increasingly isolated world, focused on older adults in care facilities.
The Association for Advancing Automation (A3) is North America’s largest automation trade association representing more than 1,400 organizations involved in robotics, artificial intelligence, machine vision & imaging, motion control & motors, and related automation technologies.