Dozens of companies are manufacturing robots that look like humans. One of them is training a machine to be a butler and plans to test it soon in people’s homes

One recent morning, I knocked on the door of a nice two-story house in Redwood City, California. Within seconds, the door was opened by a faceless robot wearing a beige jumpsuit fitted to its slim waist and long legs.

This slender humanoid greeted me with an accent that sounded Scandinavian, and I offered my hand. When our palms met, it said, “I have a firm handshake.”

When the homeowner, a Norwegian engineer named Bernt Børnich, asked for bottled water, the robot turned around, went into the kitchen, and opened the refrigerator with one hand.

Artificial intelligence already drives cars, writes essays, and even writes computer code. Now, humanoids—machines built to resemble humans and powered by AI—are about to move into our homes to help with daily chores. Børnich is the CEO and founder of a startup called 1X. Before the end of the year, his company hopes to place its robot, Neo, in more than 100 homes across Silicon Valley and beyond.

His company is among dozens planning to sell humanoids and introduce them to both homes and businesses. Investors have poured $7.2 billion into more than 50 startups since 2015, according to PitchBook, a research firm that tracks the tech industry. The humanoid frenzy hit a new peak last year, when investments exceeded $1.6 billion. That doesn’t even count the billions that Elon Musk and his electric car company, Tesla, are investing in Optimus, a humanoid they began building in 2021.

Entrepreneurs like Børnich and Musk believe humanoids will one day do much of the physical labor currently done by people, including household chores like cleaning countertops and unloading dishwashers, warehouse jobs like sorting packages, and factory work like assembling cars on production lines.

Simpler robots—such as small robotic arms and autonomous carts—have long shared the workload in warehouses and factories. Now, companies are betting that machines can perform a wider range of tasks by imitating how humans walk, bend, twist, stretch, hold, and generally get things done.

Because homes, offices, and warehouses are already built for humans, these companies argue that humanoids are better equipped to navigate the world than any other kind of robot.

The push toward humanoid labor has been growing for years, fueled by advances in both robotic hardware and AI technologies that allow robots to learn new skills quickly. Yet, these humanoids remain somewhat of a mirage.

For years, videos showing the remarkable dexterity of these machines have circulated online—but they are often being remotely guided by humans. Even simple tasks, like loading a dishwasher, are anything but simple for them.

“There are many videos out there that give a false impression of these robots,” said Ken Goldberg, a robotics professor at the University of California, Berkeley. “Although they may look human, they don’t always behave that way.”

Neo said “Hello” with a Scandinavian accent because it was being operated by a Norwegian technician in Børnich’s basement. (Ultimately, the company hopes to build call centers where dozens of technicians would provide remote support to the robots.)

The robot walked independently through the dining room and kitchen, but the technician spoke for Neo and guided its hands remotely using a virtual reality headset and two wireless control sticks. Robots are still learning to move through the world on their own—and they need a lot of help to do so. At least, for now.

‘I Saw a Level of Hardware I Didn’t Think Was Possible’

I first visited the offices of 1X in Silicon Valley nearly a year ago. When a robot named Eve entered the room, opening and closing the door, I couldn’t shake the feeling that this big-eyed robot was actually a person in disguise.

Eve moved on wheels, not legs. And yet, it still seemed human. I thought of Sleeper, the 1973 Woody Allen sci-fi comedy full of robotic butlers.

The company’s engineers had already built Neo, but it hadn’t learned to walk yet. An early version hung from the wall of the company’s lab.

In 2022, Børnich joined a Zoom call with an AI researcher named Eric Jang. They didn’t know each other. Jang, now 30, worked at a robotics lab at Google’s headquarters in Silicon Valley, while Børnich, now 42, ran a startup in Norway called Halodi Robotics.

A potential investor had asked Jang to gather information about Halodi to see if it was worth investing in. Børnich showed off the company’s humanoid, Eve. It was something he had dreamed of building since he was a teenager, inspired—like many robotics specialists—by science fiction (his personal favorite: the 1982 film Blade Runner).

Jang was fascinated by the way Eve moved. He compared the Zoom call to a scene from the sci-fi series Westworld, in which a man attends a cocktail party and is stunned to discover that everyone there is a robot.

“I saw a level of hardware I didn’t think was possible,” Jang said.

The potential investor didn’t invest in Halodi, but Jang quickly convinced Børnich to join forces.

Jang was part of a Google team that taught robots new skills using mathematical systems called neural networks, which enable robots to learn from data representing real-world tasks. After seeing Eve, Jang told Børnich they should apply the same technique to humanoids.

The result was a transatlantic company they renamed 1X. The startup, which has grown to around 200 employees, now has over $125 million in funding from investors including Tiger Global and AI startup OpenAI.

All This Is Learned Behavior

When I returned to the company’s lab about six months after meeting Eve, I was greeted by a walking Neo. It had been taught to walk entirely in the digital world. By simulating real-world physics in a video game–like environment, they could train a digital version of the robot to stand, balance, and eventually take steps.

After months of training this digital robot, they transferred everything it had learned to a physical humanoid.

If I stepped in Neo’s path, it would stop and move around me. If I pushed its chest, it would stay upright. Sometimes it stumbled or seemed unsure what to do, but it could walk around a room like a person.

“All this is learned behavior,” Jang said, as Neo’s steps clicked against the floor. “If we put it in any environment, it should know how to handle it.”

However, training a robot to perform household chores is a completely different challenge.

Because the physics of loading a dishwasher or folding clothes is extremely complex, 1X can’t teach these tasks in a virtual world—they have to collect data inside real homes.

When I visited Børnich’s house a month later, Neo began struggling with the stainless-steel refrigerator door. The robot’s Wi-Fi connection had dropped. But once the hidden technician reset the Wi-Fi, he smoothly guided the robot through its small task. Neo handed me a bottle of water.

I also saw Neo loading clothes into a washing machine, carefully squatting to lift laundry from a basket. And as Børnich and I chatted outside the kitchen, the robot began wiping down the countertops. All of this was done by remote control.

Even when controlled by humans, Neo can drop a cup or struggle to find the right angle to throw an empty bottle into the trash under the sink. Though humanoids have advanced rapidly over the past decade, they are still not as agile as humans. Neo, for example, cannot raise its arms above its head.

For the uninitiated, Neo can also be a bit unsettling—like anything that looks partly human and partly not. Talking to it is especially strange, since you’re actually speaking to a remote technician. It’s like talking to a ventriloquist’s dummy.

‘What We’re Selling Is More of a Journey Than a Destination’

By guiding Neo through household tasks, Børnich and his team can collect data—via cameras and other sensors built into the robot—showing how those tasks are performed. The engineers at 1X can then use this data to expand and refine Neo’s abilities.

Just as ChatGPT can learn to write essays by analyzing online text, a robot can learn to clean windows by identifying patterns in hours of digital video.

Most humanoids—including Musk’s Optimus and similar projects like Apptronik and Figure AI—are designed for warehouses and factories, arguing that these more controlled environments will be easier for robots to handle. But by selling humanoids for use in homes, 1X hopes to gather massive amounts of data that could ultimately teach robots how to manage the chaos of everyday life.

First, though, the company must find people willing to host an early version of a strange new technology—and pay for it.

1X has not yet set a price for these machines, which it manufactures in its own facilities in Norway. Building a humanoid like Neo costs roughly the same as building a small car: tens of thousands of dollars.

To reach its potential, Neo must capture video of what happens inside homes. In some cases, technicians will see what’s happening in real time. Essentially, it’s a robot learning on the job.

“What we’re selling is more of a journey than a destination,” Børnich said. “It’s going to be a bumpy road, but Neo will do truly useful things.”

‘We Want You to Give Us Your Data on Your Terms’

When I asked Børnich how the company would handle privacy once humanoids were inside customers’ homes, he explained that technicians, working from remote call centers, would only take control of the robot if they received the homeowner’s approval through a mobile app.

He also said that data wouldn’t be used to train new systems until at least 24 hours after collection. This would allow 1X to delete videos customers didn’t want the company to use.

“We want you to give us your data on your terms,” Børnich said.

With this data, Børnich hopes to produce a humanoid capable of performing nearly any household chore. That means Neo could one day replace workers who make a living cleaning homes.

But that’s still years away—at best. And given the growing shortage of workers who handle household cleaning and the care of the elderly and children, organizations representing these workers welcome the rise of new technologies for household tasks, as long as companies like 1X build robots that work well alongside humans.

“These tools could make some of the most exhausting and dangerous jobs easier and allow workers to focus on things that only humans can provide,” said Ai-jen Poo, President of the National Domestic Workers Alliance, which represents housekeepers, home care workers, and nannies across the country.

Soon after, Neo began cleaning the tall side windows of the house. Then, as I turned toward Børnich, I heard a crash on the kitchen floor. After an electrical malfunction, Neo had fallen backward and blacked out.

Børnich lifted the robot, as if it were a small teenager, carried it to the living room, and laid it on a chair. Even when passed out, Neo looked human.

Other humanoids I’ve encountered can be intimidating. Neo, standing about 5 feet 6 inches tall and weighing around 64 pounds, is not. Still, I wondered whether it could hurt a pet—or a child—if it fell.

Will people let this machine into their homes? How long will it take to improve its skills? Can it truly free people from their daily chores? These questions remain unanswered. But Børnich presses on.

“There are many people like me,” he said. “They’ve dreamed of having something like this in their homes since they were kids.”

---

Cade Metz is an American journalist and author specializing in technology, artificial intelligence, and Silicon Valley. He works as a reporter for The New York Times, where he covers topics related to tech companies, advances in AI, robotics, and the social impact of digital innovation. Before joining the Times, he was a correspondent for Wired, and he is the author of the book Genius Makers, which explores the history and key figures behind the rise of modern artificial intelligence. His work stands out for its analytical depth and his ability to translate complex technical concepts into accessible and insightful narratives.

The New York Times es un periódico diario estadounidense fundado en 1851 y con sede en Nueva York, considerado uno de los más influyentes y prestigiosos del mundo. (Wikipedia) Cubre noticias nacionales e internacionales, reportajes de investigación, análisis, cultura, opinión, economía, ciencia y estilo de vida. (Wikipedia) El medio opera tanto en formato impreso como digital, y ha consolidado una gran base de suscriptores online, lo que le permite ejercer un rol clave en el ecosistema mediático contemporáneo. (Wikipedia)