Spring has sprung. It sprung a bit too far forward here, as a matter of fact, as we’re wading through late-summer temperatures in New York’s beautiful Hudson Valley. Spring is a time to embrace change and take stock of the natural world (above is a photo of an eastern newt I befriended on a recent hike) – a thing we’re attempting to do in some extremely roundabout ways here in the ole Automated newsletter.
As far as change goes, new contributors deserve new formatting. The basics of Automated are staying put. We’re just adding more bang for your zero bucks. After prettying up the podcast module a couple of weeks back, we’ve added a new section called “Startup Spotlight.”
Each week, we plan to feature a short video highlighting interesting startups in the world of robotics, physical AI, and automation. I’m planning on streamlining the submission process soon, but in the meantime, if you’re interested in being featured, we can use the podcast@automate.org mailbox, using “Startup Spotlight” as a headline. This week’s features Revimo, a Boston-based firm that’s building a better robotic wheelchair.
Rebecca’s fundraising roundup is also getting its own section in the newsletter, so you can check out the latest raises with a quick glance. Then click through for her analysis on what it all means for the industry.
As far as the natural world goes, my recent conversation with The Ohio State University’s College of Engineering dean, Ayanna Howard, has me thinking about the roboticists out there who are focused on positive change. We discuss her work in healthcare, climate change, AI bias, and more.
“The reason that I was so fascinated with SnowMotes was it was helping Earth,” Howard tells me about a climate projects she worked on for NASA. “Even though it was NASA-funded, it was really thinking about how to ensure that we survive the climate apocalypse here on Earth. Working on SnowMotes was my start into thinking about Earth as a planet. And of course, who's on Earth? Earthlings? Humans?”
Is improving life the ultimate goal of robotics and automation? Is it a goal that companies, investors, or researchers can lose sight of? As RAI’s Kate Darling notes below, when it comes to elements we’re seeking to alter with automation, it never hurts to look from as many angles as possible and be open to goals that shift along with the human populations they’re intended to serve.
Vivian Sun has been around the block in autonomous driving. There were stints at trucking companies TuSimple and Waabi, where she respectively served as VP of business and chief commercial officer, followed by a fun as Amazon’s head of autonomous driving.
In March, Sun pivoted to a new — though related — field, joining Genesis AI. The move was a gamble of sorts. While still not fully mature, we’re more than a decade into the world of big self-driving bets. In spite of a seemingly endless glut of demo videos, physical AI is a nascent technology by comparison.
It can be difficult to ignore some glaring parallels with the early days of self-driving. If past is, indeed, prologue, the landscape is set to shift dramatically. It’s precisely these unknowns, Sun contends, that make the early-stage startup world so attractive.
“I really enjoyed the beginning phase, the zero to one,” says Sun. “That's why I picked the last 10 years of autonomous driving, not the next 10 years. I think this is exactly the moment we have for robotics. We're seeing true generalization capabilities. The robots can like us. There can be a pre-trained brain that can do so many things, more things than ever before. And we don't need to program for anything. You give it data, you learn, and there you go.”
Sun adds that she doesn’t expect physical AI generalization to take as long to mature as autonomous driving before it. “Obviously, we're not there yet, but we’re seeing the light at the end of the tunnel,” she says. “That's such an important spark in many people. This is the right moment for robotics, and I think the robotics industry is not going to take 10 years [to mature].”
If you’ve ever considered writing a book about technology, you’ve grappled with the fact that the publishing industry doesn’t keep pace with the tech world. It’s taken as read that certain aspects will no longer be wholly accurate by the time its release date rolls around. Of course, good technologists accurately interpret the writing on the wall long before those changes occur.
Dr. Ayanna Howard accepts these truths as we discuss some of the bigger ideas in Rebooting the Machines, six months before the book’s release.
“I talk about things like military AI and the fact that AI at some point —- maybe even by the time November comes — will be able to actually pull the trigger,” Howard notes. “What does that mean if you then still have inaccuracies in facial recognition or inaccuracies in identifying social economics? It means that when AI pulls a trigger, you might have a lot more impact depending on who you are. Those are world problems.”
Per its publisher synopsis, the book, which hits shelves November 10, grapples with the notion that, “AI, including robotics, is not built with a moral center. That means that if we want it to serve all humanity, we must provide the compass.”
Howard, who currently serves as the dean of The Ohio State University’s College of Engineering, has been engaged with such questions for the entirety of her robotics career. Her PhD thesis, “Recursive Learning for Deformable Object Manipulation” focused on the use of robot grippers to safely manage needles amid a horrifying global pandemic.
“Back in the day, we had really big issues with HIV,” Howard says. “It was on the rise, and in hospitals, the way they did waste management is that they would have waste linen, for example, would be in one bag, and you'd have just waste papers and all in another, and you would have the needles in another bag. We didn't have those safe containers that are in the hospital rooms. And people would have to basically sort them out. The idea was we get robots to do this automated management of sorting out the bags so that things that were hazardous to people, you put it in one bin, and things that had to go to wash, you put it another bin. That was the holy grail of what I was trying to do with my thesis.”
Physical AI is moving fast. But Matthew Johnson-Roberson, founding dean of Vanderbilt’s College of Connected Computing, says robotics is still missing something fundamental. The field has data, models, and momentum, but it still does not have the simple learning objective that helped language models scale so quickly. Watch on YouTube>
Sergey Levine (Physical Intelligence) - The buzzy physical AI startup is making a big bet on real-world training to kickstart robotics' data flywheel.
Martial Hebert (CMU) -The Dean of Carnegie Mellon’s School of Computer Science breaks down the university's advances in autonomy and offers insight into what the future holds.
When it comes to pieces like arms and end-effectors, plenty of robotics companies will tell you the right path is the one of least resistance. In most cases, that means off-the-shelf components — i.e., integrating parts from a third-party that has already solved that particular problem. The other common solution is building full-stack, in-house. There’s also a third way when you have the resources of a Locus Robotics: instead of just buying the components, you buy the company that makes them.
The automation pioneer this week announced that it’s picking up Nexera Robotics.
The Vancouver, BC-based firm makes NeuraGrasp, an end effort Locus is set to integrate directly into its recently-ish announced Array pick-and-place robot. “We built NeuraGrasp to solve the manipulation challenges that have held robotic picking back for years,” per Nexera CEO Roy Belak. “Joining Locus Robotics gives us the platform, scale, and customer base to bring this breakthrough technology into the high-velocity fulfillment environments it was designed for, where speed, reliability, and real-world adaptability matter most.”
The nation’s amphibians are on the move this month, migrating from winter habitats to vernal pools for breeding. The list includes various salamanders, though not the massive hellbender species, which spends all its time in water, regardless of season. The Pittsburgh startup of the same name is making money moves this month, with a $12.5 million seed round. The firm builds on the city’s storied association with autonomous systems, with a series of domestically manufactured camera systems. Pre-orders for the Tadpole, Vine, and Stereo (pictured) open next month.
In spite of making up around 0.01% of the ocean, coral reefs contain somewhere in the neighborhood of 25% of its species. Speaking of neighborhoods, these dense marine megalopolises are under threat and retreating at an alarming rate. Naturally, researchers are looking where we always look in the pages of Automated: robots. A team at the Woods Hole Oceanographic Institution (WHOI) built an autonomous underwater vehicle designed to survey these populations of marine life.
“Coral reefs are in decline worldwide and there are various ongoing monitoring, conservation, and restoration efforts happening around the world,” one researcher told A3 contributor, Liam. “We were originally motivated to find scalable solutions to monitor the biodiversity on a reef to monitor changes in reef ecology. However, the problem is intractable using current technology. This motivated us to work on the simpler problem of finding regions of high bioactivity that are correlated with biodiversity hotspots and hence makes the biodiversity monitoring problem tractable by enabling practitioners to focus on most relevant region.”
We got the extended tour while visiting Boston Dynamics’ HQ last month for our interview with Atlas’ director of product, Aya Durbin. My first extended in-person glimpse of the production version of the humanoid genuinely impressed me. It’s a high bar. I’ve been looking at a lot of robots for a lot of years — a fair number of them humanoid. Still, the system's fluid movements and strength may well put it in a class of its own, owing to some impressive feats of engineering and in-house actuators and other bits.
Right around the time we popped by, the team was apparently shooting this video of Atlas lugging around a 50-pound mini-fridge. It’s probably a coincidence that the demo dropped as college kids are leaving their dorms for the summer, but I can’t help but add moving apartments to the growing list of future seasonal jobs for humanoids. One thing Durbin and I discussed at length was why Boston Dynamics continues to invest in these sorts of flashy videos. You’ll be able to hear that full convo in our upcoming episode of the Automated podcast, but in the meantime, this bit from the blog post accompanying the video by Atlas’s director of robot behavior, Alberto Rodriguez, jumped out at me,
But less pragmatic tasks also have a purpose. For example, handstands and backflips are possible on a 90-kilogram (198-pound) robot because we have excellent thermal management, which means Atlas will be able to work in hot environments. And these behaviors train other transferable skills — how to move with agility and balance, how to use a full range of motion in constrained conditions, how to recover from slips and falls.
We did an even more recent podcast recording with Kate Darling, now of Boston’s RAI Institute. During the chat, she alludes to her recent work reframing the “three Ds” of automation — those being “dull,” “dirty,” and “dangerous.” I assume a majority of readers who have made it this far into a robotics newsletter understand the general gist, but just in case, these are generally framed as the difficult or boring jobs that are prime targets for automation, because people don’t really want to do them in the first place. RAI’s paper on the topic asks both what and, just as importantly, who defines these roles.
Darling notes, “[W]hile it might seem intuitive which tasks to put into these buckets, it turns out that there are some underlying social, economic, and cultural factors that matter. It’s possible to measure the danger of a task or job by using reported information. There are administrative records and surveys that provide numbers on occupational injury rates and hazardous risk factors. While that seems straightforward, it’s important to understand how this data was collected, reported, and verified.”
Revimo's Niko is more than a wheelchair. It aims to increase independence for people living with disabilities, transferring users without the assistance of human caregivers.
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.
.png?upscale=true&width=1120&upscale=true&name=Screenshot%202026-05-18%20at%2010.30.43%20AM%20-%20Edited%20(1).png)
