A Field Trip to the Future

The future is going to be spectacular - I know because I went on a field trip there and this is my report from the field. Our goal here at WeRobot is to figure out how to get these robots to help us at home, and I came away with such an overwhelming feeling of optimism and enthusiasm it’s hard to put it into words. I met Spot from Boston Dynamics, played pickleball with a robot, and watched a robotics vibe-coding session. I took my son, and we got a crash course in the state of robotics. Here’s the highlights from the 2025 Robotics Summit and Expo.

Very Few Companies Are Focused on the Home Humanoid

The keynote address was given by Aaron Saunders, the CTO of Boston Dynamics. You have all seen the youtube videos online from Boston Dynamics. The ones where the robot does parkour and backflips. The videos are incredibly entertaining, and I had high hopes for the keynote. Saunders, said, “How far are we from having humanoids in our home? I don’t know, maybe it’s 10 years. Maybe more.” 

Obviously I was disappointed. I’m ready for Rosie now! In fact, Saunders even had Rosie from the Jetsons on his slide deck.

He laid out the roadmap for Boston Dynamics which included a large focus on industrial robots. There are obvious advantages to the industrial focus, namely profit. Big manufacturing companies have all kinds of uses for robots to walk around and check stuff. The robots take pictures of machines to make sure they are in order, check heat settings, etc… This is a very useful application of robotics, but it doesn’t help us here in our journey for the home. They see the ‘home era’ of robots as 2035 and beyond. 

Hot take: He is wrong. More on that later. 

Sidebar: Will Robots Take Our Jobs?

Saunders also laid out that he doesn’t think robots will replace human labor. He said the labor shortage was so bad that robots are simply going to be making up the shortfall. There is projected to be an 11% labor shortage by 2030 according to the future of work report. This was fascinating to me, but that’s a topic for another week. 

What Humanoids Were There? 

There were 3 humanoid robots at the conference. Boston Dynamics makes a humanoid called Atlas, which is awesome, but they didn’t bring it, which is lame. The three humanoids were:

1. Unitree G1. This is the one we’re most familiar with and the one that I’ve been looking at purchasing. It was the demo version that has flat hands which cannot pick anything up. It’s also… short. It’s only about 4’ 2” tall. That is only 14 inches higher than a standard countertop, which means wiping the countertops is going to be challenging and putting away dishes on high shelves it not going to work. The H1, Unitree’s more premium humanoid is much more expensive but it’s 5’ 6” which feels more appropriate. The G1 can do marketing things, like dance, shake hands, and walk around. It’s very cool, but not practical for much of anything. Even though the G1 is basically a toy, Unitree's breakout session was packed, it was standing room only. There is a TON of interest in the humanoid future.

   

   Shaking hands with the Unitree G1

2. Themis from Westward Robotics. This one was pretty cool, and I had not heard of Themis yet so I was pumped for the demo. I’m excited for Themis because home use is on their roadmap which is unusual. Themis was supposed to handle a coffee service but there was an issue and it couldn’t serve coffee on day one. There seemed to be a lot of bugs with the demo, but it’s exciting early progress and we’ll keep our eye on Themis for sure.

   

   Themis from Westwood Robotics handling a tray of ‘drinks’

3. Abi from Andromeda robotics. This is an Australian company that is focused on providing companion care to older people. It was cool, but it’s not intended to ever do chores for us. This was a heartwarming application, but it’s designed to be friendly not necessarily helpful around the house.

The Four Basic Problems With Humanoids
We identified 4 basic problems in humanoid robotics right now. We’re condensing some very complex problems into generalizations but it helps to cut the problems into chunks we can analyze. Spoiler alert - number 4 is our big problem. 

1. Seeing. This is mostly solved. Robots can detect visual objects pretty well right now. Most robots use Intel RealSense cameras which are relatively inexpensive and work great according to everyone on the demo floor. This is why self driving works because the camera and the software are good enough to recognize objects in the real world. The robotics industry is just ironing out details at this point. I saw a demo where a guy vibe coded using an open source robotics library (ROS2) and Cursor and got a super simple robot to work in a few minutes. It was pretty cool. 

   

2. Ambulation. Walking around. This is also mostly solved. Boston Dynamics Spot walks around and can avoid obstacles. Unitree has robots that walk. Some robots roll on wheels. This is a hard problem, but it’s mostly solved. A lot of the demos show people pushing robots to try to make them fall over, and it's really cool to see how the robot ‘catches’ itself before falling. 

3. Thinking. Still a problem. AI has made huge strides here, but most of the AI still needs to run on the cloud. It is used extensively for training the robots, simulating things, and learning. Basically AI has helped develop the robots more quickly, but it’s still not running the robots the way it needs to. There is not really an advanced AI that lives INSIDE the robot. We need to reduce power so that an AI can run onboard and doesn’t need a human controller or a wifi connecting to a cloud-based AI engine. This is a problem, but if a robot has to be tethered to wifi, I think most of us will be okay with that. 

4. Manipulation. Picking things up, handling delicate objects, sensing pressure and adjusting force—these are what everyone kept calling “hard problems.” I kept asking: Can the robot tell if it’s holding a marshmallow or a marble? And the answer? Not really. Until a robot can move something solid, like a salt shaker or a coffee mug, and then switch gears to gently wipe the surface with a sponge or towel, it's not going to be much help wiping down the kitchen counter.

   These are the ‘Big 4’ issues in robotics, and diving into #4 is our next step.

Hands are Hard

Remember when I said I think the humanoid revolution is coming sooner than they think at Boston Dynamics? There’s cool companies working on the manipulation problem. I met the folks from Psyonic, who were awesome. The demo was given by Dale Dimassi, a film maker and marketer for Psyonic. Part of the reason Dale's demo was so authentic was because he was wearing one of the devices. As someone without an arm, he demonstrated how the Psyonic arm attached to his body allowed him to pick things up and shake hands, it was awesome. Check out a cool video of the day and life of a bionic hand user.

Not only was it emotionally inspiring, but it gave me a lot of hope that the manipulation technology is closer than we think. Pysonic is already doing trials with the robotic, touch sensitive hand being attached to robots. I hope to do a deeper dive on Pysonic at some point.

Wrap

I’m still on the hunt for a robot that can actually help with chores. After the Robotics Summit, I’ve got a much clearer picture of where things stand—and where to look next. One thing’s for sure: getting a Unitree robot without hands is like buying a car without wheels. If it can’t manipulate objects, it can’t really help me. So the search continues… until next time.

The future is bright,
Ferol