Kicking robots is back, folks. Wow. Another soon-to-be viral update from Unitree #robotics. They are beating the *crap* out of the thing and I don’t mean the kicks. I mean hitting the floor. The highlight for me is the cartwheel with the *HEAD* impact. They might’ve modeled that in the simulator.

Life hack: If you want to sound smart to your friends, instead of saying "AI" say "foundation models." And when they cock their head, say "Oh, foundation models are the specific type of AI models, like ChatGPT, that everyone's fussing about" and make a smug face. Then prepare to find new friends.

I'm firmly in the "AI is dumber than you think" camp.

Be advised. You can’t quite play DOOM on a vape pen. You may now go about your business.

Me in a recent talk: “We don’t kick our robots as much anymore to show stability. It’s increasingly seen as needlessly visually violent by the field.” The field:

In a time when humanoid gimmicks are omnipresent, this trick tickles me. A moonwalk by the KAIST humanoid from Hae-Won Park’s lab. It feels so purposeful in its use of foot contact.

24 species of fish found spawning in the Chicago River. I’m sure there’s a joke in there but I don’t know enough about Chicago to make it. Apparently it’s good news: “suggest(ing) the once severely polluted US river is now able to sustain and support resilient, biodiverse animal populations.”

This is such a consistent refrain from people trying to use chatbots for real work. It’s almost like it’s a technology built atop probabilistic inferences made from training data.

My first music composition software was Mario Paint.

I wish I was still on the island.

Good to know South Park has time management skills on par with my undergraduate students.

A fine question about domestic humanoids. But even if we assume there is a market and it can do *anything* reliably useful, my immediate concern is safety. We have no standards to establish how likely a humanoid is to fall on grandpa.

Should we be excited about the recent results from Mars? Absolutely! But we also need to understand exactly what they do, and do not, mean. 🧪🔭

Are all birds descended from dinosaurs? “They certainly are. Just like humans are a kind of primate, birds are a kind of dinosaur.” Goodbye, Mark Norell. It was always a delight to interview you. www.nytimes.com/2025/09/13/s...

The underrated pleasure of getting a timely table at a no-reservations restaurant.

┏┓ ┃┃╱╲ in ┃╱╱╲╲ this ╱╱╭╮╲╲house ▔▏┗┛▕▔ we ╱▔▔▔▔▔▔▔▔▔▔╲ make tiny extra pancakes ╱╱┏┳┓╭╮┏┳┓ ╲╲ ▔▏┗┻┛┃┃┗┻┛▕▔ specifically for the dog

I’ll be playing the diabolical social deduction game of “Blood on the Clocktower” with some friends, starting at 1pm ET. Should be a fun one, and I’ll be in the chat! youtu.be/Q-SXzfVzSJg?...

Evan Ackerman has been a thoughtful reporter on robotics for a decade or more @spectrum.ieee.org spectrum.ieee.org/humanoid-rob...

LOL. For those who don’t know, “vibe coding” is a posh term for just letting ChatGPT do all your coding for you. I am shocked (*SHOCKED I SAY*) that actual programmers need to swoop in after to redo it.

In Nov last year, we published a predictive theory of locomotor adaptation to novel environments, capturing learning and generalization phenomena in over ten experiments: www.nature.com/articles/s41... I'm honored to be featured on the 2025 Innovators Under 35 list for this work. ter.li/7c0jek

Being a professor is so weird. You are constantly having to act like you know things.

Can we save drafts on this website yet? I had a nice long thread I was drafting (videos and images galore) and it got blown to smithereens when I changed apps.

Slow news day otherwise, but make sure to check out the cool NASA announcement about past life on Mars: www.cnn.com/2025/09/10/s...

The latest edition of Ask Dr. Hubicki is up! Topics include the environmental impact of AI, the statistics of playing Plinko, and whatever the heck a "mondegreen" is. youtu.be/LDyznTZyioY?...

The NSF Integrative Movement Sciences Institute (IMSI) is a major biology research center spanning over 20 US institutions. The Mission: To understand why we move like we do--from molecules to muscles to our environment. At our Summer Institute, we train new students in our labs--like Hannah Chow:

A huge thanks to @chubicki.bsky.social for leveling up our science communication game at @movementscience.bsky.social !!

now THAT'S what I call "low-poly"

People are dunking on this clunky Tesla Optimus demo — I get it. In a weird way I’m happy(?) they released this instead of something staged. Two very practical lessons: 1. Conversation delay kills user experience 2. Being escorted by a robot that can’t walk at my pace would drive me to violence

Me: “It’ll be fun to livestream the @movementscience.bsky.social Summer Institute Symposium to YouTube. I’ll just pack some equipment in my luggage.” Some equipment:

How fast do robots think? How fast do they *need* to think? One of the ironies of current robots is their reaction times are so much faster than people and animal and yet we’re still way better at moving and reacting to the world.

Acrobatics are always a crowd pleaser. Nifty use of deep reinforcement learning by the Robotics and AI Institute (spinoff of Boston Dynamics). What will get me really excited is to see this in the wild, jumping off rocks and rough terrain.

Today’s the annual Capstone Symposium for the Integrative Movement Sciences Institute. Students from across the country are presenting their cutting-edge research into how humans and animals move. As outreach director, I’m streaming the whole event live! Tune in now! youtube.com/live/bR2AmMs...

Good news about the cutting-edge AI tools from the $3-trillion-dollar company.

I’m back at #dragoncon this year! I’m giving three talks: FRI@10: “Robot Science or Robot Fiction?” The latest on the reality of robots in the movies. SAT@530: “This Year in Humanoid Robots 2025” SUN@1130: “The Math that makes Robots Think” A fun and easy romp through the math of robot control

Nice *three* consecutive backflips on Spot. Boston Dynamics has really taken to using deep reinforcement learning in the past year. Love the agility. Would love even more for another worldwide legged robotics challenge to push the autonomy. youtu.be/LMPxtcEgtds?...

Inside Tesla’s facilities, engineers are training their robots by recording people constantly doing menial tasks. Now they’ve switched from wearing mocap suits to head-worn camera rigs. I was asked for comment along with some great robotics colleagues. www.businessinsider.com/tesla-musk-o...

The Raibert hoppers were a triumph in the nascent days of legged robotics. Walking was too hard… so why not run first? A clip from today’s coding stream with @shiffman.lol from Coding Train: www.youtube.com/live/3FeySIv...

A beautiful exploration of how to visualize and understand slide puzzles from a literal higher dimension. youtu.be/YGLNyHd2w10?...

Robotics Coding Livestream Imminent! I’ll be joining the fantastic @shiffman.lol of Coding Train to talk about how to make robots hop and run. Can we make a simulation of a hopping robot before we get too hungry for lunch? Find out starting at 11am ET today! www.youtube.com/live/3FeySIv...

It’s that time of the year again! Do you have a question for @chubicki.bsky.social? Get your questions in before we record next week: robhasawebsite.com/askdrhubicki

Great demo from Boston Dynamics/Toyota Research Institute. Other robots do manipulation (hand tasks) while people mess with them, but mostly while standing upright and still. Here it steps and crouches to pick up parts, combining locomotion with manipulation quite nicely. youtu.be/HYwekersccY?...

A major challenge for future humanoids on space missions: fixing themselves. Robots break all the time-- and the more jobs they have, the more ways they break. Robots will have to take care of themselves before they can take care of a Mars habitat. --- Podcast clip with youtube.com/@ellieinspace

What’s the deal with all these humanoid robots? Are they crewing a mission to Mars? (no) I had a fun chat with Ellie in Space all about it. We get into why there’s a boom in humanoid robots, their limitations, and their prospects for space travel and working in our world. youtu.be/TfeVv7yQYlk?...

Assembling furniture with no help gives me this same irrational pride and I have a PhD in robotics and mechanical engineering.

Oh there's absolutely a Korok Seed there 😄

Horrible no-good very bad idea. OpenAI is buying its own hype thinking that slop-prone word factories are a tool for everything. And they’re gifting that tool to the federal government. www.cnbc.com/2025/08/06/o...

“Clanker” is apparently a newly emerging slur against robots — and it offends me. I have a bedrock moral principle against giving Star Wars references any more cultural impact. www.npr.org/2025/08/06/n...

No human advance is more victimized by its own success than vaccines.

A one *billion* dollar offer? For *one* researcher to join your AI project? Which they TURNED DOWN? This might be a bubble, folks. www.wired.com/story/mark-z...

Wow. Roboticists have long wondered who the law will blame in a self-driving car crash. We’d joke: “engineers won’t decide the fate of the technology, lawyers will.” Drivers will blame automakers and automakers will blame drivers. Yesterday, a jury blamed both. Driver: $86.5M Tesla: $242.5M