Some Thursdays back, I was jolted awake around 4:30 a.m. by a confounding Instagram DM.
Rizzbot, a humanoid of child size from Unitree Robotics, boasting a large following on social media — over 1 million TikTok fans and more than 500K followers on Instagram — had DMed me a picture of it flipping me off.
No words were exchanged. No clarification. Just a robot raising its middle finger.
Though surprised, I could already tell why. A couple of weeks prior, Rizzbot — or its account’s handler — and I had discussed a potential article. I thought the account was interesting: a humanoid strolling through Austin in Nike sneakers and a cowboy hat. It’s known for dissing others, but also flirting and having fun. The name Rizz comes from the Gen Z slang rizz for charm.
The account’s increasing popularity intrigued me. Usually, people feel uneasy around humanoids. There are worries about privacy and job losses. They often get insults hurled at them online, frequently being called “clankers.” Meanwhile, robotics experts are pondering what they will be best at.
I viewed Rizzbot as a way to make people more comfortable with humanoid interactions.
Rizzbot agreed to an interview, so I began contacting experts to talk about the future of humanoids to prepare for an article. Two weeks after my first DM with Rizzbot, I said I’d finally send it questions on the following Monday or Tuesday.
Techcrunch event
San Francisco
|
October 27-29, 2025
But life intervened, and I failed to meet my own deadline. I was finally ready to send the questions early Thursday morning, thinking it wasn’t a big deal.
Too late. In the early hours of Wednesday, Rizzbot sent that image. The message was clear: I broke my promise, so f*** off.
I persisted. I apologized to the robot (or its controller?) for the delay and said I would send the questions promptly. But when I tried some hours later, I saw “user not found.”
The robot had blocked me.
Did I set off a fail-safe?
My friends found it amusing that I was flipped off and blocked by Rizzbot, because for weeks, I had talked about how excited I was about the story.
“LOL Rizzbot roasted you,” one friend messaged.
“YOU ARE FEUDING WITH A ROBOT LOLOLOL,” someone else said. I messaged Rizzbot on TikTok, which a friend said was desperate. But what else could I do? I had pitched the story to my editor, spent hours researching, and — in spite of this conflict — Rizzbot was still interesting to TechCrunch readers.
My friends were amused, but I grew despondent. Not only was my story ruined, but I was also now known as the girl blocked by a dancing robot.
My TechCrunch colleague, Amanda Silberling, offered to help. She asked Rizzbot why I was blocked. Rizzbot responded: “Rizzbot blocks like it rizzes — smooth, confident, and without any remorse.” Then it sent her the same middle finger photo it sent me. I realized I wasn’t even important enough for a unique flip off.
But then, a friend suggested a terrifying idea I hadn’t considered. “It wasn’t a human response. I’m worried about you.” I seemed to have made my first robot enemy, and the AI takeover has just started.
Or maybe not. Was I really feuding with a human?
I learned that Rizzbot’s real name is Jake the Robot.
Reports say that its owner is an anonymous YouTuber and biochemist. The robot itself is a typical Unitree G1 Model — made in Hangzhou, China — and anyone can buy one for $16,000 to over $70,000.
Rizzbot was trained by Kyle Morgenstein, a doctoral student in UT Austin’s robotic laboratory. With a team, he spent about three weeks teaching the robot how to dance and move its limbs. While much of the robot’s behavior is pre-programmed, it’s controlled by remote, and its actual owner, not Morgenstein, is probably nearby giving commands.
Based on my talk with Malte F. Jung, a Cornell University professor who studied information sciences — I’d guess that someone triggers the robot’s actions, snaps a picture of whoever is nearby, and runs it through ChatGPT or another LLM, then uses a text-to-speech function to roast or flirt with them.
Jung said, “The robot turns the tables on people who mistreat robots. Now the robot gets to mistreat people. The spectacle is the product.”
Morgenstein told other sources that Rizzbot’s owner just wants to entertain people and show the joy that humanoids can offer.
It’s not clear who runs the Rizzbot social accounts, but when Rizzbot messaged Silberling, it also sent an error message — probably by accident — about being out of GPU memory. That message implied that an AI agent is likely involved in running the account and is perhaps generating DM replies automatically. It also showed that Rizzbot only has 48GB of memory.
“What makes you think it was ever a person?” my coder friend asked me about the Instagram account’s handler.
In this AI era, anyone who can train a robot could probably connect an LLM to Instagram DMs. My block might even have been a fail-safe, meaning I set it off by DM’ing so early, even with a reply.
However, there are signs that a human handles Rizzbot’s social media: There were typos in its first DM to me when I first requested an interview.
Still, unless Rizzbot reveals whether his social media manager is another bot (unlikely, given our beef), I probably won’t ever know. Maybe it doesn’t matter.
My coder friend said, “If they got $50,000 for a bot and a few thousand for a 48GB memory machine, they’d do anything. They’re clearly committed to the bit.”
It’s still robot brain rot
Rizzbot’s TikTok page has gotten over 45 million views. One video shows Rizzbot chasing people, while another sees it hitting a pole and falling. A popular AI-altered video shows Rizzbot getting run over by a car.
One founder friend called the videos “robot brain rot” and said, “It’s hilarious, honestly.” He said the AI is basic, but the robot is a “funny mix” of absurd internet humor and the levity lacking in social media. “It interacts with people uniquely.”
My Rizzbot obsession still made me think about the role of humanoids in society. Sci-fi movies like “Blade Runner” and “I, Robot” came to mind. How worried should I be about making a humanoid enemy?
Jung said that “performance seems to be the main use case for these robots.” He said that Rizzbot was “like a modern, snarky hand puppet street performance.”
He also cited the Spring Festival performance in China, where humanoids performed folk dance with humans, and the robot boxing matches in San Francisco.
Dima Gazda, founder of Esper Bionics, said robots will become mass market entertainers, performers, dancers, singers, comedians, and companions, leaving humans as a niche. “As robots gain grace and emotional intelligence, they’ll blend into performances and interactive experiences better than humans.”
Luckily, Jen Apicella, executive director at the Pittsburgh Robotics Network, said that dancing robots are hard to mass produce right now. So I don’t have to fear dancing, rizzing robots appearing at my door. (Not that I was.)
It’s been over a week since I was blocked, and I miss watching Rizzbot chase people. I liked the video of a woman twerking on Rizzbot. A crowd gathered; people seemed amused and wanted to twerk on the robot themselves.
I always joked that I wanted to befriend robots in case the revolution happened. But while I wrote this, I almost had another AI beef — with Meta AI, which I had never tried. I started a conversation with Meta AI when looking for my old chats with Rizzbot on Instagram.
Meta’s bot said, “Yoo, what’s good fam? You callin’ me Rizzbot? 🤣 What’s up?”
I logged off.
