My forecast: the most valuable use of humanoids isn't cheap labour but expensive art. Not a loader but a performer. Not a worker on the assembly line but a celebrity millions come to see.
At the weekend, at half-time in the Brazil — Norway match at the World Cup, it wasn't a person who walked out onto the pitch of the New York stadium. It was a robot. Boston Dynamics' Atlas walked across the field, recreated the goal celebrations of Haaland, Kane and Son — the very goals the whole world rewatches on a loop — and handed the ball to the referee to start the second half. Tens of millions of people in the stands and at their screens weren't looking at a factory. They were looking at a show.
Remember this picture. It isn't about football. It's about where humanoid robots are really heading. And it's exactly what I've been saying for a long time.
People will tell me: a one-off stunt, an ad for Hyundai. Quite the opposite. Hyundai bought out Boston Dynamics entirely and is building a factory in Georgia for Atlas — 30,000 robots a year by 2028. This isn't a one-night attraction. It's the first public appearance of a mass-market product. And the first thing that product did in front of the whole world wasn't to assemble a part on a conveyor — it danced Haaland's goal.
When people say "humanoid robot", almost everyone pictures one thing: it'll come to the warehouse, stand at the machine, haul boxes and mop floors. Cheap labour that replaces a loader for a couple of dollars an hour.
I'm convinced this is a mistake. And here's why.
A factory doesn't like the human form. Two arms, two legs, a head, a hinge at the waist — that's evolution's compromise for the task "survive in the savannah", not "assemble 500 parts an hour with no defects". Where you need speed and precision on a line, a specialised machine will always win: a robotic arm that makes one movement perfectly and for ever; an assembly line built for one product. It's cheaper, faster and more reliable than any humanlike robot. A humanoid on an assembly line is like harnessing a ballerina to haul sacks.
The humanoid is the most expensive and most complex form of robot. Spending it on what a simple manipulator does more cheaply is economic foolishness.
The answer is simple: the human form is for people. Not for parts — for us.
For millions of years we've learned to read another person: how they move, keep a rhythm, celebrate, look you in the eye. The humanoid is the only form of robot that hits this built-in "receiver" of ours. We can't admire a manipulator. But we freeze when a humanlike being does a somersault, dances to music or recreates a favourite footballer's celebration.
Note a detail almost everyone missed. Atlas was taught these movements through retargeting — a technology that transfers a living person's movement onto a robot's body. The engineers literally filmed how Haaland celebrates and "put it on" the machine. All of a humanoid's expensive insides — balancing, reinforcement learning, thousands of simulations — are tuned for one thing: to move like a human in order to move humans. It isn't a work tool. It's a tool of attention.
Let's do the maths — I'm an entrepreneur, not a dreamer.
A loader robot competes with a person who costs little. Its ceiling of value is the wage of the worker it replaced. Putting a $30,000 humanoid there doesn't pay: it takes years to pay back and still loses to a narrow machine.
Now take the performer robot. It competes not with a loader but with a celebrity. Atlas's one appearance at the World Cup was seen by tens of millions of people — that's an advertising contact brands pay millions of dollars for. Unitree's dancing robots at the Chinese TV New Year gala drew an audience the biggest stars could only dream of. A robot like that pays back even its highest price — not because it saves on wages but because it produces attention. And attention today is the most expensive commodity on the planet.
A factory robot saves pennies. A celebrity robot earns millions on attention. That's the difference between a cost and an asset.
The whole idea in two lines:
— As a loader, a robot is worth as much as it saves.
— As a performer, a robot is worth as much attention as it gathers.
Attention scales to millions of viewers. Mopping floors doesn't.
I'm not fantasising. While everyone discusses when the robot will come to the warehouse, robots have already stepped out into the world — precisely as performers:
Here's my list of professions where I'll see the humanoid first: musician, host, showman, coach in dance, yoga and martial arts, children's entertainer, stage performer. All of this is work not with parts but with people. Work for attention and emotion.
People will tell me: MILAGPT doesn't make metal robots, it makes software AI assistants. What does a dancing Atlas have to do with it?
Everything — because the direction is the same. I'm not building MILAGPT as "cheap labour that will replace a person for less". I'm building AI partners and assistants for people: a nanobot friend that remembers your life and speaks your language; a business worker that looks after the client and replies around the clock. Their value isn't that they're cheaper than an employee. Their value is in relationships and attention: they're always there and always attentive to the person.
It's exactly the same fork as with humanoids. You can build a robot as a cost — to save money. Or you can build it as an asset — to create value through a connection with people. I always choose the second. In hardware and in software alike, the future belongs to robots that work for people, not in place of cheap labour.
The robot of the future isn't a silent loader in a workshop. It's a performer, a partner, a conversation companion, a coach. Someone people come to for attention, emotion and mastery. The factory will go to narrow machines. The stage, the hall and the human heart — to the humanoid.
My forecast in one sentence: the first truly mass-market humanoid will be not a worker but a star. It will step not onto the assembly line but onto the stage. And it will earn not by being cheaper than a person, but because millions of people will want to watch it.
That's exactly why I build both metal robots and our digital nanobots by one law: value through a relationship with a person, not through cheap labour. Let simple machines do the grinding. The humanoid was made for something bigger.
Forecasts, launches, the release of the MILA token, new articles and invitations to take part — on the founder's channel.
MILAGPT · Tashkent
Want a digital employee like this for yourself?
Write to Mila: she replies in seconds, at any time of day or night.
Write to Mila →Tell her about your business in Telegram and she'll help from the very first conversation.
Get your own Mila →Share:Telegram