Posted by josefchen 20 hours ago
1) Unitree robots are largely made by hand.
2) Labor is approximately 8% of cost of manufacturing these robots. Because these costs are not prohibitive, there is little incentive to automate.
3) Labor costs are higher in America.
4) Therefore, American companies have higher incentive to automate.
5) Therefore, the West will capture most of the robotics opportunity.
Arguments 1-4 are reasonable, but 5 is a non-sequitur. Higher incentive to automate does not imply a better ability to automate & scale manufacturing.
I believe Unitree is currently assembling robots by hand for practical reasons:
1) Labor is still relatively cheap.
2) Humanoids and robots are evolving rapidly, changing every few months.
3) Custom assembly lines are expensive, and become obsolete when the product changes. It's difficult to maintain an assembly line when the product is rapidly evolving.
4) Tooling is inadequate. Some robotic parts are delicate, complicated, and difficult to automate manufacturing of. American companies have hit the exact same problem: Tesla is assembling the hand/forearm of Optimus by hand because it hasn't figured out how to automate manufacturing them. https://www.theverge.com/tech/1016492/tesla-optimus-robots-p... . Apparently, Tesla has hundreds of handless Optimus robots just standing around, waiting for hand manufacturing to catch up.
5) Demand is still relatively niche, these are not mass market products. Unitree only sold 5,500 robots last year. There is not enough demand to justify massive automated lines.
The author also ignores the biggest counterpoint: the Chinese have already demonstrated mastery in automated manufacturing, and have proven they can run dark factories (factories without humans) at a better scale than the west.
When humanoid robots become a mass market product (they are not ready today because the software side is majorly lacking), Chinese manufacturers are in an excellent position to rapidly automate.
Motor winding is an amusing problem. Everybody making motors has a powered winder. It's handling the wire ends that can be either manual or automatic. If you look at videos of winders, automating the wire end handling is most of the machinery.
Here's an overview of winding automation:
Full auto, robotic transfer.[1]
Full auto, manual transfer [2].
Semi auto. [3].
Manual, powered. [4]
Manual, unpowered motor overhaul.[5] (Country girl in China who fixes motors and engines.)
Those winding machines are all from the same company, NIDE. You can buy as much automation as you want, depending on your labor costs. Option [4] is cost-effective if you can have one machine tender per machine.
[1] https://www.youtube.com/watch?v=7FiCQX0y7-A
[2] https://www.youtube.com/watch?v=3wZvSem4Cd8
[3] https://www.youtube.com/watch?v=G4vqLxMlvk0
1. The terrible quality of the locally manufactured items. I am in Sydney and will absolutely not buy anything marked Made In Australia. I have been burnt too many times.
2. The nearly 100% mortality rate among the companies.
No, that is only one customer. There need to be lots of customers for a company to survive. So, ideally you recommend, based on good quality.
You take a bit of a gamble if you need warranty or customer service later on.
Some sort of network effect again, I suppose. There's a better term that's can't find now.
Just to extend your point - chinese car manufacturers are not only producing more cars, they're innovating at a faster pace too. While I'm not an expert, there is enough and more on social media around this. Eg: https://x.com/dhh/status/2113500613082329576?a=fscIadYr4U1xm...
The key parts are R&D, design, intellectual property, engineering talent. The question is, how well are Chinese organizations doing in those capacities in regard to robotics. My impression is that the answer is, quite well.
The focus on manufacturing is (in many cases) a twisted approach of economics. In the mid-20th century the West did very well with mass production. It was an innovative technology at the top of the value chain. But in early-to-mid 21st century, it's (often) an old, ~commodity technology and the ROI is poor. It's also physically not good for workers in many cases. When looking for economic growth you want to aim for top of the value change, for the future, not what was done 70 years ago - it won't work the same way economically anymore. In economics, you can't return to what worked before.
But the economy is in equilibrium which means as this decreases, it's going to change the value of other things, and open the door for new things. Whoever can take what LLMs offer, and then most successfully build on top of that will probably be 'the winner.' I'm not sure exactly what this will translate to in practice, but it seems reasonable to predict that manufacturing is going to be a huge part of it.
"But as the cost of 'cognitive work' in general starts approach $0, the value of such, in and of itself, is also going to approach $0."
When the pirate party was big in EU, there was much talk about copying approaching the cost of 0, and therefor we would never pay for copying data, only creating new data.
I guess this all due to copyright law, and it will not disappear anytime soon, but still. I would not be too quick to draw that same conclusion.
Imagine all the work involved in, for instance, trigonometry and its basic applications in the past. And the later on with the advent of kinematics and so forth. These things involved a lot of computation and so there were literal human calculators whose job it was to work out all of the calculations.
But with the advent of electrical calculators, we literally no longer know that these jobs even existed. It's very easy to see the same in a future where, for instance, computer programs can be efficiently written from description alone. At that point the idea of a 'programmer' would seem like quite an odd concept.
I'd also look at what it did to the value of such work. Now computations, even quite complex, are mostly worth exactly $0. But we built extensive systems on top of those computations that became worth quite a lot indeed!
However, we do have Wikipedia. Wikipedia was never valued at the peak sales price of Encyclopedia Britannica times the number of people online, if it was it would be worth around $16 trillion.
But it's difficult to leap ahead from step 0 to step 3, as you need the gains from step 0 to achieve step 1, the benefits from step 1 to achieve step 2, etc. Some of those benefits are food, shelter, healthcare, education (hard to have engineers without it, for example), physical infrastructure, finance, honest and skilled government, etc.
Yes, absolutely. You can see many places have followed that path; you can skip the economic policy and technological R&D and just use mature solutions - South Korea and Taiwan, for example.
The 'cheating' - I assume you mean about IP - is nonsense to me. Who made the rules? The people who want to preserve their IP, and they would make different rules if they didn't have IP themselves. China didn't make those rules.
The argument is that Unitree's assembly process is currently very manual because there is no incentive to automate, which is because labor cost isn't the bottleneck here, where in the US labor cost are much higher, therefore the US will have incentive to automate more. Those are reasonable so far, until you try to use that as a conclusion that this is why the West will win?
First of all, what does he think the other 92% of the cost of manufacturing is? A small portion of that is materials and electricity, but the biggest part of the remaining cost is assembly tooling/fixtures for the assembly line, and a fully automated assembly line is much, much more expensive compared to a semi-automated or manual production line.
In other words: it's not that Unitree couldn't build fully automated assembly lines if they wanted to, but spending tens of millions of dollars fully automated assembly line that takes up floor space and requires even more labor for manual maintenance for a product that is only being produced by the hundreds to thousands in quantity every month rather than millions simply does not make economic sense to begin with.
Second, why is he assuming that a fully automated assembly line will be cheaper to build in the US than in China, especially since he toured a factory with fully automated assembly lines at BYD in China immediately after? And sure, I would absolutely argue that American software engineering is more efficient than Chinese when looked at as a whole, but the cost of programming an assembly line is a minor portion of the cost compared to actually getting the assembly line fixtures and tooling built.
Third, his cost estimate is way off if he thinks you can get the same product manufactured in the US for only 50% more. From my own anecdotal experience, a more realistic cost for US manufacturing for pretty much anything is minimum 300% to 1000% more expensive if you are to make the in the US, and the lack of a mature parts supply chain in the US is honestly a bigger issue than worker wage.
If instead just creating few of the tools, Destin pushed to create $1T worth of it, then maybe the supplychain will automatically build around it.
If China's labor cost risen to the level that makes Unitree robots more appealing, then Unitree will figure out ways to automate their production line more to keep up with the demand.
All and all, it's not a capability problem, it's a "why do it, and why now" problem. Just like, say, sending people to Mars. The tech is already here, but why send people to a planet that (currently) don't even have enough stuff needed to support living? Except of course, Mars will probably stay this way much longer than a country changing course.
Also, just a sidenote: I think Destin misread the market. He wanted to create a "high-quality", "durable" product. Sure, it make sense to do it most of the time, but a brush? You know, that thing gets dirty in an dark and oily way very quickly. With a durable brush that's expensive to replace, then you have will a problem of maintenance, clean it etc. And grill oil is not easy to clean, do you now also have to buy another product just to maintain the brush? How about now I offer you a cheap made in China brush that will disintegrate in your hands after 5 times of use, but then you just throw it away and get a new one, no maintenance required. Of course, throwing away too many trash is bad for the environment, but guess how the other trash gets handled, including, say, toothbrush?
How about, invent a grill that don't need to be brushed? Just a rinse and it's done? Just a Smarter way to think about inventing, you know.
If I were to design it to be manufactured in the states, I would probably go for a simple wooden handle with off the shelf wire mesh instead of chainmail, and the fastening would just be mechanical fit between wrapped wire and handle with no additional fastener, which should massively improve price and manufacturability.
“”” America has run this experiment before. In 1964 about half a million Mexican farmworkers were shut out. In 1965, 20% of California's tomatoes were picked by machine; by 1970 the figure was 99.9%.“””
I honestly had not considered closing down (legal) immigration as a forcing function for automation. Industrial policy as immigration law.
But I don’t think I get the rest …
Higher costs onshore mean machines give us an advantage - i get that. The rest I may be missing
5 years ago I was trying to get a custom aluminium enclosure for some circuit boards manufactured. We found potential suppliers in Taiwan, but they basically would have needed an expert onsite to be able to set up all the jigs. They also could have supported us through rapid scaling by switching to moulded raw stock which then was post machined, and would bring down unit prices a lot.
However, one guy knew a local SoCal machine shop that supplied parts to the auto modding scene and had a bunch of 5-axis milling machines. For 200 units they would beat Taiwan on every metric: cost, time, effort on our side, flexibility for changes.
We're engineers. From this lens, the difference between us and cavemen are the tools we have. We should lean into this.
So industrial policy should be focused on high end engineering quality, more than bringing jobs back
(Might be spotting a political problem there but in general I think I agree)
People are much more likely to try something if they see peers and role models doing it. Even better if it’s something fun and exciting that they want to be a part of.
I think to convince someone to run a factory instead of trying to become MTS @ lab, they should know 1. it’s not a binary “make it or break it” where failure means doing some boring, poorly paid job (the downside is good but not so guaranteed that it’s abused/attracts the wrong people ) 2. it’s interesting and challenging and enjoyable at times 3. it’s meaningful at a societal or personal level. That's what made tech/AI/finance work and it's the competing pitch.
Why should someone with other options want to work in or operate a factory?
I'm seeing this in India. Some really smart students are doing some fascinating startups at home because emigrating to the west has become really hard.
Still a very small effect, but one that gives me a lot of hope and cause for excitement.
(Again apropos of nothing I am assuming the rise of “on the vine tomatoes” in packets is presumably because cutting four off the branch is quicker and easier than twisting four or five.
And even possibly there is a robot doing it. Would be interested to know
Defense contractors charge maintenance fees both domestic and abroad. For abroad, the maintenance contract is a requirement of getting the deal passed through congress and is frequently more expensive than the gear itself.
SaaS charge monthly instead of one upfront fee for a well tested product that used to be the norm
Video games similarly have been transitioning to day 1 paid DLCs and monthly fees
Even the movie industry is a subscription now (streaming providers aside). There's not one movie, there's the MCU pumping out 3 movies a year for the last 20. How many movies from last year have a number in their name?
Then there's of course, the actual rents. And insurance. And taxes. Something has to give, but I fear it will be the humans.
(I'm not affiliated with the site in any way; I just found it useful and thought I'd share it here.)
The conference venues have concrete flooring, so I just suddenly hear loud rhythmic thumping as these things stomp around a corner.
After the initial "huh that's a cool robot thing" thought, it really started giving me the heebeejeebees.
Uncanny Valley is an evolutionary advantage, I’ve surmised
Awkward to think about why tho
Probably too many pinch points though.
> the robots that make robots will be Western.
is quite reminiscent of the dismissive attitude of many tech people here towards stuff that doesn't originate in the US. It really feels strange to have lived through this change, from a place that used to be pretty cosmopolitan and focused on tech, to a place one can't browse without reading daily jingoistic comments.
It also feels like an atmosphere change rather than a deliberate one, in that people say that stuff casually, almost without thinking about it or the implications thereof. These ideas have permetated so much into the discourse that people don't really scrutinize them before repeating them.
Ok, but what drives the cost of the parts? I am sure the parts manufacturers spend on labor, and the parts the part makers buy depend on labor. Labor drives all vast majority of the expense if you go all the way down the supply chain.
If we automate more and more of that labor with robots, then it drives down all the costs.
If we get truly autonomous robots that have the mobility and precision of humans, they could replace all of the labor. The 'switching' cost goes way down compared to the machines that do the work today.
I’ve really become fascinated by the hypothesis that outlawing slavery supercharged industrialization, and that the global crusade against slavery is why we had this whole technology explosion.
Rome knew about steam power. There’s some tantalizing evidence some ancient cultures played with electricity and maybe even simple forms of electric lights. There’s Greek stories about automatons. But why try to build machines to do real work? Just have a slave do it.
If we’d abolished slavery in ancient Roman times maybe we’d be looking at photos of habitable exoplanets from probes arriving in the Tau Ceti system by now.
Tocqueville wrote a bit about this in Democracy in America. Despite the South having an edge in terms of "free" labor the institution of slavery greatly suffocated human potential to such an extent that the non-Southern states still had higher productivity.
What do you have in mind here? The Dendera light is a carving of a lotus.
The non-woowoo proposals I’ve read call for very primitive apparati that would have been possible with metallurgy of the time. If some, I dunno, bored priests and scribes over thousands of years figured out some neat tricks. It’s before the big medieval alchemy craze but thousands of years is a long time.
I also forgot about early optics, the Antikythera mechanism which could not have been a one off, early use of fossil fuel, etc.
I don’t think the thesis I referenced argues that abolition is the only thing that made tech take off, just that it supplied an economic forcing function missing for most of the rest of history.
"But the demand for robotics is being driven by government agencies and research institutions, rather than everyday consumers, for whom the products are still too expensive and not yet practical for home use. Humanoid robotics and robotic dogs have been popular in the entertainment industry but are yet to be useful in home settings.
Shares in Unitree, China’s leading robot-maker, have halved since they started trading in August."*