Where’s my robot butler?

ABC Listen, podcasts, radio, news, music, and more.

I'm Rachel Brown, and in all my time reporting on crime and courts for the ABC, I've seen a lot of monsters. But this crime chilled me more than the others.

The man behind it, a pilot you wouldn't think twice about if you met him on the street. Someone who flies under the radar.

Someone with secrets simmering just below the surface.

Huntsman, the latest season of the ABC's true crime podcast, Unraveled. Here Unravel Now on the ABC Listen app.
This podcast was produced on the lands of the Wabakal and Gadigal people.

Every so often a video pops up online that makes it feel like the future has arrived. My name is Bern, and today we're launching Neo, our humanoid for the whole.

Tech company OneX's robot NEO was launched at the end of October 2025 with an incredibly slick-looking promotional video. Neo is a humanoid companion designed to transform your life at home.

It combines AI and advanced hardware to help with daily chores and bring intelligence into your everyday life.

The robot is about the size of an average 14-year-old, but it seems much more eager to do household chores than the average 14-year-old. We built the chores feature to give you your time back.

You give your Neo a list of chores, you schedule a time that you want them done so that you can focus on what matters to you while your Neo does the rest. Every parent's dream.

And for the first five minutes of the video, it seems amazing. Hey, I'm Neo.
I'm here to help around the house. But the actual reality of what's happening is somewhat less amazing.

If there are any chores that your Neo hasn't learned how to do autonomously, you can use Expert Mode to have an expert from 1X supervise the session and provide corrective intervention to help Neo complete any task.

So, Neo will do any chores it's capable of doing autonomously on its own. But if there's a chore it can't do, an expert from the tech company will dial into into the robot to operate it remotely.

So sometimes Neo is an autonomous robot and sometimes Neo is just a very, very expensive puppet. Eventually, Neo will become fully autonomous, capable of helping you with anything around the house.

But how many tasks can it currently do autonomously? When the Wall Street Journal sent a reporter to find out, they discovered that the answer is almost none.

To be clear, on my visit, I didn't see Neo do anything autonomously. Neo was clumsy and awkward in every motion.
It also seems to have broken its finger while loading the dishwasher.

And that took five minutes.

But despite this, several big tech companies like OneX and Tesla are currently pouring billions of dollars into building autonomous humanoid robots, the robot butlers and maids, straight out of science fiction.

And people like Elon Musk are convinced they're going to change the world. I think humanoid robots will be the biggest product ever.
The demand will be insatiable.

For decades we've been obsessed with this idea of a far-flung future where we all have hovering cars and live-in robot maids to do our chores.

But even after all this investment, robots still require an enormous amount of human intervention to do even the simplest tasks. So why has creating a functioning humanoid proven so difficult?

And why are we still so willing to fall for the hype? I'm Matt Bevan and this is If You're Listening.

We have been imagining robot domestic servants for a very long time. Rolo, answer the door.

Yes, Rolo the robot, the chromium-plated butler is just a daydream after all.

People were obsessed with the idea that robots would soon make housework a thing of the past, and a little bit annoyed that they weren't already available.

When those robots finally did arrive, people also seemed to hope that they would be kind of dishy.

This is Tomorrow Control.

Here you'll witness the creation of the amazing Miss Animation, Phase 1, the body.

This exhibit toured the world in 1968 under various names like Miss Animation, Roberta the Robot, and Miss Honeywell, and it purported to show a real robot made.

The body houses a world of scientific marvels, a busy network of light cells and mercury relays.

The future Tronic computer, with its flashing lights, similar to a human heartbeat, taps out the message that Paul is A-O-K.

In each demonstration of the robot, a headless robot body would be put inside a cabinet with a hole in the top. The doors would be closed, and then

this is the transistorized head.

A robot head will be put into the top hole, then a plastic cover will be attached, covering the head, then the doors would open.

The robot body and head have somehow become attached, and out walks a robot. Hello, it's a here-to-be pleasure.
It sometimes takes her speech computers a while to warm up.

The completed quote-unquote robot would would step out and start doing household chores. Now, the robot looks interesting.

She's wearing a black skivvy, a silver mini dress, and has a silver head complete with silver speed dealer sunglasses. And she walks very much like a human doing an impression of a robot.

You and our housewife of the future will be showing us what push-button living in the future will be like. However, not too many buttons, because many actions will be pre-programmed.

So when I first saw this with my cynical, sceptical, jaded 21st century eyeballs, it was obvious that this was a trick. Not a particularly good one either.

A quick search of US patent applications shows exactly how it works and how the mannequin body and head are inserted into the cabinet and then switched with a woman who is wearing a robot costume.

But it's clear from newspaper coverage at the time that many people who saw the exhibition were fooled by it.

Miss Honeywell toured France, the UK, the US and Australia and the international media had a field day.

Some newspapers reported on the incredible robot while other reporters tried to expose what they described as a hoax.

British papers claimed they saw the robot breathing. Australian papers said that the woman acting as the robot had thicker legs than the mannequin and that's what gave it away.

One reporter from the Adelaide Advertiser actually grabbed the robot's hand and said it felt like warm, feminine flesh. He was then pounced on by security guards.

When she wasn't being sexually harassed by tabloid journalists, Miss Honeywell's performance involved vacuuming, reading books and presenting the fantastic scientific advancement that can do anything, the ultimate appliance.

This one appliance combines all household tasks into one centralized, completely automatic unit.

So the ultimate appliance was basically just a large box adhered to the wall that could apparently deal with everything from dirty clothes to dirty dishes.

But watch, as Anna places them inside the ultimate appliance and then adjusts the dials for wash, sterilize, and dry.

When asked if any of this was real, the owner of the Miss Honeywell illusion generally just said that he would leave it up to audiences to decide.

And while some journalists did expose the hoax, many people did believe that robots like Miss Honeywell were going to be available at some point soon.

Which is why a decade later, when another questionable robot started a global tour, people were happy to believe that that one was real too.

Meet Klaytu.

Billed as the world's first domestic android.

Klaytu may be just a mixture of batteries, circuits and remote program animators to you, but he's supposed to do anything your real-life gentleman's gentleman will do.

And what's more, this one's guaranteed not to get into the cooking sherry. Klaytu the robot arrived in Australia in 1978 as a guest of Australian Women's Weekly for their 45th anniversary.

He looked like a less threatening Dalek, and its creator Tony Rickelt claimed that it was at the bleeding edge of technology. Tony Rickelt has developed Klaytu over a 10-year period.

He hopes to have them on the market in the USA by 1980. He says the domestic Klaytu will do most domestic chores and have a vocabulary of about 250 words.

In front of the Sydney media and Women's Weekly editor Ida Butros, Tony Rickelt proceeded to demonstrate Klaytu's capabilities. Klaytu, would you please open up your audio circuits?

Open audio circuits. He then opened Klaytu up to the media for questioning.
Klaytu, what do you think of Australia?

Australia

is a lovely place.

Klaytu then proceeded to do kind of a robo-stand-up set. Have you ever thought of going into the radio business?

Radio.

I was in love with a radio one.

How long did this affair last?

Then Klaytu and Ita Buttrose cut the cake. Can you cut the cake, Klaytu?

With assistance.

I mean, putting a knife in the hands of an apparently autonomous robot is an interesting move, Ita. The media was quite taken with this robot that could cut cakes and crack wise.

In the short term, the next development is to give Klaytu flexible hands with fingers. But even in his current state, Klaytu is a remarkable machine.

Riquelt Riquelt said that Klaytu was just the beginning.

We're working on a domestic android right now for a special order for an archaeologist in New York that'll have 125,000 word vocabulary, be capable of speaking fluent English and fluent Spanish simultaneously.

That type of robot is about $75,000. That's like $375,000 today.

It wasn't only the ABC and Women's Weekly reporting on Klaytu's potential. Riquelt managed to get industry publications to report on it too.

Interface Age magazine ran a gushing article on Riquelt titled A Dream That Came True about the company's 32 robots that traveled around the world doing demonstrations, always accompanied by their technicians.

Of course, the technicians were actually operating Klaytu via a remote control and providing his voice via a hidden microphone.

It was an elaborate and bizarre hoax, bordering on a fraud, really, and Rickhelt and his company seemed to have vanished without a trace around 1980.

And honestly, there is a fascinating lesson in these stories for us in the media.

Miss Honeywell and Claytoo were clearly not real robots, and yet journalists covering them were hesitant to say that outright.

Even the ones who expressed skepticism relied on critical quotes from experts rather than just telling their readers or viewers that the robot is obviously fake.

Most journalists seemed eager to believe the hype and were much less interested in the actual robotics work that was being done by actual scientists in actual labs at the time.

This is Shaky the Robot.

While Miss Honeywell and Klaytu were dazzling journalists all over the world, scientists at Stanford University were teaching a robot to navigate between obstacles by bonking into them.

Now what's hard to keep in mind here is that there are no human beings involved.

To get Shaky to do all these things by remote control would be a pushover, but Shaky is making all these decisions by himself.

The Stanford researchers also would move objects around to try and confuse the robot. The intermediate intelligence also allows him to recover from unforeseen events.

Despite the fact that this was an actual autonomous robot with no humans driving it, the mainstream media was not particularly interested.

Shakey ended up in science magazines, Klaytu ended up on the TV news. Through the 80s and 90s, American interest in humanoid robots ebbed a bit.

And by the time I was a kid, all of the most exciting work in robotics was being done in Japan.

This is a distant relative of C3PO. It's been developed by Honda Motors.
Researchers have taught it to climb stairs and predict it will soon be a part of everyday life.

In 2002, Honda released Asimo, a humanoid robot which looked like an eight-year-old in a spacesuit. My name is Asimo.
Nice to meet you.

Thanks to the newest technology developed by a human being, I can do a lot of things just like an ordinary person. But like all the humanoids before him, Asimo wasn't autonomous.

He couldn't make any decisions or solve any problems. He was a very cool-looking, remote-controlled robot.
May I ask you to take a photo with me?

Honda spent more than 30 years and an undisclosed number of of millions of dollars on Azimo before eventually retiring the program in 2018. They had never sold a single robot.

So, why?

Why have all of these attempts to build humanoid domestic helper robots failed?

Well, it's worth noting that despite our inability to build a robot in the shape of a human, Robots in other less shapely shapes are having incredible success in all sorts of other places.

Of course, we've had robots of a sort for years. For more than 50 years, we've had autonomous trains, automated warehouses and logistics and industrial robots.

Japan has half a million of them working in its factories. And there's a few key reasons why those simpler robots have worked while domestic humanoid robots have not.

Reason number one is that in the case of robots assembling cars and driving trains and forklifts, we've gone to great efforts to make sure that there's no humans getting in their way.

Humans are generally not allowed anywhere near these robots while they're operating. Reason number two is that those robots are specifically designed to do a single task.

They don't need to like attach an axle to a Toyota and then go make a cup of tea and then stack things in the dishwasher.

Reason number three is that none of these robots are trying to walk around on two legs.

Human bipedalism is the result of literally billions of years of trial and error through evolution, and keeping us upright requires thousands of tiny reflex movements every single second.

If you imagine walking on uneven ground, these reflexes help you balance quickly, before you even know it, by correcting small changes. And that helps you avoid a fall.

For two-legged robots, this is spectacularly complicated, which is why they fall over all the time.

And why they seem to struggle with even very simple tasks like folding laundry or loading dishwashers. It's incredibly difficult to build a human-shaped robot.

And yet, according to Elon Musk, for a domestic robot to be useful, that's the shape it's got to be.

Well, if you wanted to do all the things that a human can do, it turns out you need a human-bodied robot. We built the world to suit humans, so we need human-shaped robots to operate in it.

It turns out humans evolved to the shape and capabilities that we have

for good reasons. That's true, but not really relevant to this discussion.

Robots don't need to go through the same evolutionary development process as humans because the world we live in now and the tasks that we want robots to do for us are entirely different to the ones that we as humans needed to do thousands of years ago.

We evolved to be this shape with these arms and these hands and these legs so that we can run and throw things at wild animals at the same time, not so that we can fold laundry and load dishwashers.

Take for example the one domestic domestic robot that has broken into mainstream usage, the Robo Vacuum Cleaner. This is the future of domestic cleanliness, the world's first robotic vacuum cleaner.

Instead of building a robot that could operate a traditional vacuum cleaner, we designed a robot that was a vacuum cleaner.

But despite this, in May 2025, Elon Musk posted a video of one of his Optimus robots sweeping up a mess with a dustpen and brush, and then using a handheld vacuum to clean the floor.

When Musk posts AI-generated videos of potential future tasks being done by Optimus, they're using power tools designed for humans, delivering first aid in the way a human would, and mysteriously going to a casino and betting on poker and roulette.

It would surely be cheaper and easier to design robots that do things in a robot way, rather than forcing them to do things in a human way.

And even Elon Musk says that it would be easier to have several robots designed to do specific jobs rather than one robot that can do everything a human can do. So if you want to just do a subset,

that's much easier.

I often think about C3PO and R2-D2 from Star Wars. R2-D2, a pleasure to meet you.
I am C3PO Human Cyborg Relations.

C-3PO, a humanoid robot, struggled to do anything except shuffle around and complain. R2-D2, where are you? At At last, where have you been? Meanwhile, R2-D2 did all the actual work.

He was essentially a Swiss Army knife on three wheels with specialized tools for every task.

He was actually useful, like a robot vacuum cleaner. And yet Elon Musk is trying to find a use for C-3PO.
This is madness.

The story of humanoid robots is massively hyped by the media.

Everyone thinks they want a robot made or a robot butler, and so everything that resembles one gets enormous media coverage, even if it's a fairly obvious hoax.

After decades of development and billions of dollars worth of investment, we're still at the stage where humanoid robots need supervision by an operator to do even the most simple tasks.

Now, you may be wondering, won't AI fix this problem? If ChatGPT can write emails, poems, and computer code for me, surely it can figure out how to fold a shirt.

But the problem is ChatGPT learned to do those things by consuming the entire contents of the internet.

It's easy to learn a language when you have millions of emails, poems and programs to read through.

There isn't any data sets online for how to manipulate two humanoid arms and hands into shapes that result in a shirt being folded.

Which is why 1X, the maker of the NEO robot we started with, says that they're going to spend the next year shipping thousands of their robots into people's homes so NEO can create its own data sets.

We're shipping more than 10,000 units next year. And the year after, we're hoping to get to like about 80-ish thousand.

So NEO is going to spend its downtime observing its owners and gathering a huge amount of data on them, like a child watching its parents talk and move.

And what we have here is the ability to live and learn among people and actually capture all of those values. So I think that data advantage is huge.
So at best, it's a robot toddler.

At worst, it's a puppet that also spies on you. In 2026, if you buy this product, it is because you're okay with that social contract.
If we don't have your data, we can't make the product better.

So is Bernt, the founder of 1X, and Elon Musk, really all that different from Tony Rickelt? Basically, just a glorified robot grifter?

At the moment, these tech bros are trying to find customers for a product that doesn't work yet, and they're promising that it will sometime in the future.

It's been more than 50 years since people around the world were tricked by Klaytu, a wise cracking fiberglass cone-shaped puppet, into thinking that robot butlers were just around the corner.

It turns out we aren't really all that much closer now than we were back then.

If you're listening is written by me, Matt Bevan. It's produced by Adair Shepard.
Supervising producer is Cara Jensen-McKinnon.

If you've come across any other stories about questionable or strange robot projects, let us know. We'll be talking more about them next week.

You can send us an email or a voice note question to ifyou'relistening at abc.net.au. Catch you on Tuesday.