Robotics and humanoid robotics & civilian drones discussion

sunnymaxi

sunnymaxi

Major
Registered Member
tacoburger said:
And they're going to stay a developing country for much longer if they keep chasing the bandwagon and not investing early in obvious techologny
Click to expand...
completely wrong.

do you even know, how much money Chinese have spending in 4th industrial revolution related techs. like Software , autonomous driving , EV supply chain , 5G/6G , Ai , Quantum , green energy , supercomputers , IOT and digitalization.

tacoburger said:
For companies like Xiaomi or a commerical consumer level robot companu like Unitree? The former would have had the funds to invest in a side project like that even 10 years ago and the latter would have to invest an obivous area like human like robots, it's like being a internet company and not investing in software development. Both of them found the funds to invest in robot dogs afterall.
Click to expand...
wrong again.

Unitree founded in 2016. i can even tell you the exact day and date when they received first proper funding.. neither Xiaomi's first robotic dog project started 10 years ago.
 
T

tacoburger

Junior Member
Registered Member
sunnymaxi said:
do you even know, how much money Chinese have spending in 4th industrial revolution related techs. like Software , autonomous driving , EV supply chain , 5G/6G , Ai , Quantum , green energy , supercomputers , IOT and digitalization.
Click to expand...
And they are big movers in the field, but almost always the 2nd comers. And like it or not, being first does matter. It doesn't matter if you're product is better, if you came out a day later then your competitor, you're automatically seen as inferior, especially with the existing stereotype of chinese companies being massive IP thieves.

Yes that matters. Even today despite the Unitree robots dogs being way more common and better and cheaper than Spot, I still see many people who go "Isn't this just a Spot ripoff" and it has been 6 years. Baidu could release a LLM that was a thousand times better then GTP-4, based on a completely new A.I architecture and people will still call it a ChatGTP-copy. It's just human nature to naturally value the first to do anything. How many of you can remember the name of the 2nd man to have ever stepped foot on the moon compared to Neil Armstrong? I often forget. What about the 3rd or the 4th? I can remember Yuri Gagarin easily but I have no idea who was the 2nd man to ever reach space, nor do I really care to find out.

And yeah it matters when you want people to buy your products and to invest in your companies.
sunnymaxi said:
Unitree founded in 2016. i can even tell you the exact day and date when they received first proper funding.. neither Xiaomi's first robotic dog project started 10 years ago.
Click to expand...
Spot was unveiled in 2016. Other robot dogs/robot horse prototypes came out way back in the early 2000s. And again, if they had the funds to sink into a robot dog wayyy back then, why not have invested in a way more potentially useful human robot wayyy back then too? Not in the last 2 years where it seems like there's a new humanoid robotics company every month. There's a reason why BD is a decade ahead of basically every other robotics company out there, it's because they started 15 years ago and focused hard on consistent R&D. Imagne if Unitree had a human robot with 5 years of R&D behind it that was unveiled to the world during 2021, instead of the rush job that looks like they started development on it last year.
 
tphuang

tphuang

Lieutenant General
Staff member
Super Moderator
VIP Professional
Registered Member
tacoburger said:

Xpeng has their own humanoid robot. And this brings up another discussion, I really think China as more of a follower and not a trendsetter. This isn't about innovation, or not to say that China didn't set off some trends of her own but it really feels like Chinese companies don't do something until some famous western company has a product that goes viral or is wildly successful, then they all jump on the bandwagon to develop their own version. This problem isn't unique to China of course, but it is a big issue.

For example, there have been half a dozen chinese companies unveiling humanoid robots this year, most of them appear to be very similar in capabilities, frankly most of them are quite bad, not appearing more advanced than the humanoid robots of 10 years ago despite the vast improvements in A.I and computing power, though some of their hands look quite impressive and most of them seem to have been in development for a short amount of time. This is not to say that Chinese scientist and engineers don't do good work and of course they will improve with time, but rather I suspect that most of this companies only started work on a humanoid robot after Tesla announced that they were working on a humanoid robot, that or after the video of Boston Dynamic's Altas robot doing parkour went viral. Both events happened during 2021, which fits most of this robotics only having around 2 years of development, you can tell that most of them are in the early stages of development and most are nowhere near being an actual product or finalized version

I don't have any hard evidence but the length of time this robots seem to be in development, the rought state of development that suggests that they are in very early stages of R&D, the fact that so many companies unveiled their projects within the half a year really suggests that yeah, a lot of this companies followed Tesla's lead, espically with car companies like Xpeng jumping in. Now I'm not saying that all of them followed Tesla or Boston Dynamic, some of them probably have been development long before Tesla announced their robot but I think companies like Xiaomi, Unitree or Xpeng only started development because Tesla was also doing so, or when Boston Dynamic's Altas started go viral.

I mean, I rather that they jumped on the bandwagon then completely ignore the development of a exciting new techologny. But why wait until Tesla? Why couldn't have some mega company like Xiaomi or some specialized robotics company not started serious development 10 years ago? This isn't even a niche tech like say resuable rockets, humanoid robots have been a goal since the early days of science ficition. Boston Dynamics doesn't have some secret sauce, they are the leaders in this field because they started serious development on humanoid robotics 15 years ago.

Another example would be ChatGPT. Despite the transformer architecture being a popular topic of research and GTP-3 proving it's potential a year before and the large investments in A.I and computing power in China, Chinese tech companies didn't really start serious development and intergration of LLMs until ChatGTP took the world by storm for some reason.
Click to expand...

that's because you are only looking at areas already with Western products.

go take a look at what Huawei developed recently or in EV world or batteries or any of the renewable companies, nothing western equivalent

you know why? because only china gets hard start, it wins

the only area you see competition are ones where Western companies got in their first

But realistically, why would you even look at XPeng? They are a joke of a company, run by a loser of CEO

Agibot design is way ahead here

Unitree actually already has its dog robot in usage

I can go further into why China's humanoid robot research didn't go into overdrive until recently, but only if you have specific questions. This seems more of a complaint post from you rather than a real thoughtful point of discussion.
 
solarz

solarz

Brigadier
tacoburger said:
Humanoid robots work well because we designed our modern world to be easily accessible to humans. Yeah treads, 4 legged or wheeled robots would be better design wise up until they have to climb a flight of stairs, open doors or walk on narrow walkways. And they're new and still in development, as their A.I and capabilities improve, their uses are basically infinite. You'll be surprised at how many low skilled jobs involve someone being there just to do things like flip a switch, turn a valve, tighten a few screws. One idea I have for a use for a human robot could be something as simple as ensuring all the doors and windows are closed, all the air-con are shut down, switching off the lights after closing hours.. I say that because it used to be part of my job, just running around making sure all the windows were shut, turning off the lights and air-conditioning in the building as we closed for the night, a moneky could have done that part of my job if we could train it not to shit on the floor.
Click to expand...

The problem with this line of thought is that tasks which humans find to be very simple are extremely difficult to program. Right now, these robots are only demonstrating an ability to walk around and perform simple manual operations. There's a long way to go before they can become as versatile as you imagine them to be in the post above.

For example, just for the simple task of flipping a light switch, the robot needs to be able to recognize various kinds light switch, recognize that lights are turned on or off, be able to navigate the area and avoid various obstacles, both static (such as furniture) and moving (such as people or pets), and be able to understand human instructions and update it's task routine accordingly.

Think about how a human learns to do a job. They need training from another human, and this requires then to understand human speech and be able to process visual demonstrations.

All of these are extremely difficult software problems, and until you solve them, humanoid robots would not be of much use.
 
E

Eventine

Junior Member
Registered Member
tacoburger said:
I call bullshit. Humanoid robots are not a new technology. They just gained a lot more attention over the last 2 years with all the hype around Tesla and Atlas. They are also plenty examples of robots in mythology, aka the "golem", artificial humans and robots in sci-fiction, they are not a new concept. 15 years ago there were walking robots. Humanoid robotics would be quite high on the top of the most promising techologny to develop, especially when combined with the recent developments in batteries, computer vision, A.I etc etc that can boost a robot's capabilites vastly.
Click to expand...
Humanoid robots aren't a new technology, but they certainly aren't an established product, either.

Nobody has been successful selling humanoid robots for the decades they've been available in an academic sense. Well, aside from very primitive robots you might find at an attraction park.

Western tech. companies have what are called "moon shot" departments that look at technologies 5, 10, even 20 years out. Purely dedicated to long-term research & development. Basically cost centers that are funded by the profits drawn from the rest of the company, all for the potential of break-outs. The industry literally makes jokes about this - see, for example,
Please, Log in or Register to view URLs content!
, and these aren't even Google's real "moon shot" projects, which are handled by a separate department altogether that used to be called Google X.

The vast majority of Chinese tech. companies did not have such departments for most of the last two decades. They were consumer companies that focused on gaining market share via creating low margin products. Xiaomi is an obvious example here as they basically marketed themselves as a low cost Apple. R&D was not their focus because they didn't even have the money for it.

R&D is a moat that you begin to build once you become an established company, and if Chinese companies still aren't doing it today then I would agree with you. But it seems the smarter companies are starting to make that shift and investing in R&D, and as for the ones that aren't doing it... Well, they'll deserve what's coming to them.
 
B

Biscuits

Major
Registered Member
tacoburger said:

Xpeng has their own humanoid robot. And this brings up another discussion, I really think China as more of a follower and not a trendsetter. This isn't about innovation, or not to say that China didn't set off some trends of her own but it really feels like Chinese companies don't do something until some famous western company has a product that goes viral or is wildly successful, then they all jump on the bandwagon to develop their own version. This problem isn't unique to China of course, but it is a big issue.

For example, there have been half a dozen chinese companies unveiling humanoid robots this year, most of them appear to be very similar in capabilities, frankly most of them are quite bad, not appearing more advanced than the humanoid robots of 10 years ago despite the vast improvements in A.I and computing power, though some of their hands look quite impressive and most of them seem to have been in development for a short amount of time. This is not to say that Chinese scientist and engineers don't do good work and of course they will improve with time, but rather I suspect that most of this companies only started work on a humanoid robot after Tesla announced that they were working on a humanoid robot, that or after the video of Boston Dynamic's Altas robot doing parkour went viral. Both events happened during 2021, which fits most of this robotics only having around 2 years of development, you can tell that most of them are in the early stages of development and most are nowhere near being an actual product or finalized version

I don't have any hard evidence but the length of time this robots seem to be in development, the rought state of development that suggests that they are in very early stages of R&D, the fact that so many companies unveiled their projects within the half a year really suggests that yeah, a lot of this companies followed Tesla's lead, espically with car companies like Xpeng jumping in. Now I'm not saying that all of them followed Tesla or Boston Dynamic, some of them probably have been development long before Tesla announced their robot but I think companies like Xiaomi, Unitree or Xpeng only started development because Tesla was also doing so, or when Boston Dynamic's Altas started go viral.

I mean, I rather that they jumped on the bandwagon then completely ignore the development of a exciting new techologny. But why wait until Tesla? Why couldn't have some mega company like Xiaomi or some specialized robotics company not started serious development 10 years ago? This isn't even a niche tech like say resuable rockets, humanoid robots have been a goal since the early days of science ficition. Boston Dynamics doesn't have some secret sauce, they are the leaders in this field because they started serious development on humanoid robotics 15 years ago.

Another example would be ChatGPT. Despite the transformer architecture being a popular topic of research and GTP-3 proving it's potential a year before and the large investments in A.I and computing power in China, Chinese tech companies didn't really start serious development and intergration of LLMs until ChatGTP took the world by storm for some reason.
Click to expand...
Even having the largest economy in the world doesn't provide infinite resources.

China focuses on areas that are marketable and/or has important uses according to the government. For example, in the area of industrial robotics, commercial drones of all sizes, battery and material science, it is US which is the follower and China that is the trend setter.

What is marketable and desirable is different in different regions. Even though both US and China have "complete" supply chains, they excel more in different areas, and those areas automatically becomes sectors where they produce more products than in other, more neglected, areas.

Private equity, whether American or Chinese, always take the path of least resistance and maximum profit.
 
You must log in or register to reply here.
Top