It's a compliment to Unitree that when I first looked at this video with the latest updates to the G1 Bionic humanoid robot, I wondered if it was rendered and not real life. But it is real, this is what they are capable of, and the base model is only $16,000.
There are many humanoid robots in development, but the Unitree G1 Bionic is interesting because of its very cheap price point. Open source robotic development AI is rapidly advancing the capability of robots. Meanwhile, with chat GPT type AI on board we will easily be able to talk to them.
How far away are we from a world where you can purchase a humanoid robot that will be capable of doing most types of unskilled work with little training? It can't be very many years away now when you look at this.
The latest updates to Unitree's $16,000 humanoid robot show us how close we are to a world filled with humanoid robots.
byu/lughnasadh inFuturology
18 Comments
This will be a liability nightmare. The first time someone sends this to the shops to get a candy bar and it totals a car or kicks a toddler, the courts are gonna light up like a pinball table trying to allocate blame, with no precedent about AI.
This is definitely rendered, and not actual video of the robots.
> It can’t be very many years away now when you look at this.
its decades away.
its a bit counter-intuitive but robotic / hardware is much much harder than AI
It’s very cool, but in terms of cheap AI labor arriving fast I’m more excited about what [the Roborock Z70](https://youtu.be/9l-358S-CxA?si=_vjxhfJLQcRGE-Rm) signals – this is the direct sequel to a Roborock I already own, and if this trend continues, in 5-7 years something the size of a toddler could clean and tidy a significant amount of your house for the cost of one month’s rent. This seems like a more realistic entry point for “robot butlers” in the next few years, and with less liability: Fewer people are gonna try sending it to the shops. Or making love to it.
My one was impressive a couple years ago because it vacuums, mops, empties it’s dustbin, replaces it’s water and cleans and dries it’s own mop, but… It uses discrete systems for each of those tasks, so there’s a lot of potential points of failure, and human intervention is needed to clean it’s filters, etc. A few more generations and all that could be achieved by the onboard limb(s). Heck, there’s one that [can pick up other tools from its caddy as needed](https://youtu.be/cbfufgWwSc0?si=IwiW6kXlui2r7a55), and I had the hilarious image of one holding a tiny vacuum cleaner.
What kind of psychopaths programmed these androids to not give a fuck about only stepping on the white part of pedestrian crossings?
Pure madness.
I mean, that’s great, and walking around reliably is an important stepping stone, but what about picking up stuff and putting them in where they should be put, according to natural language instructions? Can they do that seemlessly? To me, this would be the bare minimum for people actually taking a mortgage to get one. If it can take out the trash and pick up stuff on the ground to group them into categories, it gets very near the point where I would consider putting that money to buy them. If all it does is walk around, I won’t pay that much.
>I wondered if it was rendered and not real life.
… Why?
Selling these robots to ordinary consumers is not on their roadmap. Today’s AI and robots are still a long way from consumer-grade products – they have short battery life, are not flexible enough, lack power, and have no suitable software. AI that can generate text, pictures and videos cannot cope with various emergencies in the physical world. Therefore, Unitree chose to sell them to scientific research institutions, universities and enterprises to jointly explore the possibilities of future robotics technology. They released videos not to market products, but to expand brand influence to attract more investment. According to an interview with the CEO, it will still take more than 10 years for robots that can replace people to do anything
Uhhh, pay $16k to watch a human-shaped machine walk around? I think I’ll just buy a new motorcycle, again.
I’d say within ten years
people won’t purchase anything.
You’ll lose your job to one of these things
People will treat you the way you treat the homeless.
No one will give you UBI because they don’t believe in “socialism”
And you’ll probably suffer and die.
Can you imagine one of these busting into your local restaurant and deporting the cooks on behalf of dear leader?
I guess ICE agents will do more meaningful work after losing their job.
The 16k price is a joke – does not include hands or compute and has an 8-month warranty lol. Essentially a paperweight. The real price is “Contact sales” [https://www.unitree.com/g1](https://www.unitree.com/g1)
– Weight: 35kg
– Battery: 200Wh
– Battery Life: 2 hours
Bull-fucking-shit.
This could be a valid help in the incoming geriatric world. Cleaning the house is mostly a pick and place operation as well as loading the dishwasher or washing machine.
I need to see one of those folding the laundry.
Someone said remote control? You will be able to hire someone without the physical presence.
Farming? Picking greens? Mowing the lawn? All these remote workers will feed the AI, until necessary. Robots will auto assemble other robots in the production line.
Can someone explain to me why people think we need them? It will never be good at everything, that’s why we have specialized robots. Food delivery rovers, roombas, self driving cars, robotic warehouse platforms and forklifts, conveyor+robot arm automated factories- all kinds of robots that are good at what they do. Even at warfare they are useless.
I’d be happy to know what’s the purpose and usage. I can’t come up with anything besides being a crutch for retrofitting factories made for humans, for couple decades until those are fully automated.
Humanoid robots aren’t as useful as we think they are, they’re a fun project, but also complex and difficult to maintain. You can buy one for 16k, great, then you need the supporting infrastructure to be right… oh shit that is expensive!
I’ve recently been to a workshop identifying what uses there might be for robots in healthcare, particularly in a hospital setting, and we’d be much better off with the sort of robots you find in large manufacturing facilities. Sorting robots, autonomous transport buggies, large scale storage facilities, robots that identify whether the linen is actually washed or whether it needs to be rejected for another run…
Not because they would do a better job than humans, but because it is increasingly more difficult to find enough humans (willing) to do those jobs. Then when you sit down and analyse what is required to operate even those ‘single task’ robots, say a fleet of autonomous delivery and collection robots and a state of the art sorting and storage facility that ‘feeds’ them, the total cost very rapidly starts to spiral: Physical space, maintenance, specialist engineers… reliable and secure network. That’s before you get to other types such as cleaning robots.
Our hospital was originally designed for the use of electric (human operated) tugs, so our main corridors are capable of dealing with that, we even have a facilities garage where they are kept, charged and maintained. But even then the pathing, signage, schematics etc. require lots of alterations to the existing fabric of the building once you start utilising autonomous vehicles.
Just an afternoon of back-of-the-napkin calculations based on conversations with various directors and leaders in the hospital, using simulated use-cases it became very clear, very rapidly that it is possible, but also requires a complete financial overhaul of the system.
Bipedal humanoid is a logistically inferior form factor for these. Why are they pushing this?
For $16k it makes me wonder why I haven’t seen them waiting tables at restaurants yet.