Semi-analysis on robots; Eric Zhao on sampling the answer space; Ethan Mollick's experiment teaming AI with an individual; Shlomo Klapper on AI and humility
Reporting from Wuhan... Robots. I've been in the factories. Robots are building robots. Auto plants are essentially all robotic with human handlers walking around should something go wrong, which doesn't seem to happen. Xiaomi now has a fully automated robotic factory cranking out their cell phones (that are darn nice BTW) at the rate of 60 per minute...one phone every second. No humans involved. Xiaomi also has their auto mfg. nearly all robotized, the cars are beautiful, and are so price competitive there's a year wait to get their new SU7.
Couple all that into supply chain logistics where everything, and I do mean everything, is next door, all feeding into rail spurs going directly to computer automated loading and unloading gantries with related dockage along the Yangtze all the way from Chongqing to Shanghai, with most of it located in the middle to lower reaches in the Yangtze River Valley Development Zone...that's just the Yangtze. The same thing exists down in the Pearl River Zone. It is truly the factory for the world. I could go on, but won't.
America is more than flat footed; it's chopped off at the ankles. There's not going to be catch up. Some gains here and there, but right now, it's looking like a horizon job...i.e., in a sailing race when one boat is over the horizon and the rest of the fleet is tacking around trying to find a favorable breeze.
It's not possible to describe "normal folk in China" due to vast socio-economic differences in societal structure between the US and the PRC. "Normal" isn't a useful adjective. The middle class is doing fine and enjoying life. The majority of citizens are not middle class; life is hard. Like really fxxxing hard. There are HUGE disparities and inequities between urban and rural citizens.. If you're rural, which is several hundred million people, you're still shut out from the economic advances. It's a wildly diverse and diffused economic picture, one that is not portrayed in the West in any coherent manner. Direct comparisons between here and the US do not lend themselves to understanding what's going on.
The robot part...is freaky. Unless the American robot makers are keeping everything secret and under wraps so we can't know reality, China is way ahead of the US in robot tech and AI integration. Just one little metric.... Robot taxis aren't something anyone wonders about. They're all over the place in Wuhan and work perfectly.
China’s has a demographic problem just like S Korea and Japan, just a decade or so behind. Then there’s a financial train wreck that is coming for them.
China is also importing a lot of oil and food which could be disrupted if they get feisty about Taiwan.
Maybe, but Western media dwells on what makes China look tenuous or awful. Financial train wreck...they're handling it remarkably well. Demographic time bomb...we'll see. Lots of Econ PhD's talk about this as inescapable doom. Consider the source.
They will get feisty about Taiwan. No if; when. There's a reason Bid Daddy has an "ironclad" partnership with Vlad. Oil, gas. Food? You need to visit China and take a few train rides cross country. Chinese dig in and endure. America is the place that freaks out, not China.
In short, don't read FP, The Economist, or any Western media and imagine you know what's going on in China.
I have been to China a few times working at refineries and was underwhelmed by the operations. The problem I saw was silos and deference to superiors that stifled production and operations.
Peter is a bozo. He's got his community, he presents well, speaks authoritatively, etc., etc., but he weaves together a bunch of decontextualized nonsense that fits his already determined narrative. A few times in China is not enough times. When those times are matters a lot. Anything more than 5 years ago isn't relevant. I can't comment on the silos because I don't know where they're at or when they were developed.
I'm not cheerleading for China. I've been in country for 15 years. I'm just reporting facts. The differences in everything in just the last 5 years are overwhelming. Of course, anything can happen. There's evidence for everything. I can list a few disasters, but I can also "what about" any negative predictions/observations for as long as anyone wants to listen.
And really...if you're seriously listening to Pete and imagining he knows the future...you really need to find other information sources.
I don't recommend myself. I occasionally cite specific things that I see with my own eyes. If you want to know, read a LOT of stuff. There's a lot of it in Substack. China Thought Express leans negative on China, but not Peter Zeihan predicting the future type negative. Pekingnology is just plain good info. Nathan Whitaker has his Rural China Substack Blog. Robert Wu is excellent, very fair minded Chinese guy in the Mainland that reports reality both sides. The Monitor often has good stuff. There's a guy that just goes by Bill that I just stumbled upon that spent years in China doing business and he's got some really good insights into how stuff works here. Bill... Do a search for Bill with China stuff. I read Western media too because it's important to see all sides, although most of it is simpleton biased decontextualized nonsense.
The main point I'd make is don't read one thing and imagine it's the single reality. China is a very large and very complicated place and it's real easy to fall into ideological bias. It's not all one thing.
Also...keeping it simple...South China Morning Post. And, seriously, Xinhua, the government mouthpiece. Yes, there's the obvious propagandistic stuff, but it's full of useful news, as in real stuff that's happening now.
Lots of essays in Substack. I was searching and finding more than I feel like linking to.
Great links, especially from Semi-analysis and Klapper. The first because of my career in supply chain/logistics and continuing interest in technology impact there. The second because Klapper's essay had a certain profoundness to it that correctly, I believe, articulates a paradox of knowledge and faith. Acknowledging doubts requires honesty with oneself, and can lead to a more nuanced and mature knowledge - or faith. And as I look around lately, honesty is something we need more of.
Great links! And comments. If China is so leading in robotics, then it’s very very good that Trump put some tariffs on their outputs. And more would have been better. We car makers in Slovakia better push the EU to stop Chinese mfg dominance in robots and everything else, especially cars. Robots making robots will hugely reduce the value of most menial human labor. Why pay low wages to an illegal when one can buy a robot to do it?
AI not being allowed to say “I don’t know” seems the essence of the hallucination problem. Spock’s willingness to say “approximately 4.3 seconds “ was a charming nod to humility yet preciseness. Klapper’s post, combined Zhao on searching thru more possible right answers, reinforces my idea that error detection and correction remains a key problem not yet solved. But the checking of differences from different ai model answers seems likely to avoid the avoidable untruths. Likely helped hugely with a human SME (subject matter expert) to focus fact checking on likely untruths.
Mollick’s great team-AI work implies mostly, maybe exclusively, two person teams. So now I’m wondering about multi-AI teams with just one human. Or maybe two for big creativity.
Underdiscussed is the difference in deciding what goals should be sought vs the best way to achieve the chosen goal. Humans will remain most important in choosing the goal, while ai will better optimize how to achieve it. As ai navigation now does in replacing map reading with clear step by step directions, to which a driver must pay somewhat close attention to. And so many drivers are now very weak at reading and understanding maps.
Balderdash. What was it Berra said? Something like predictions are different fficult, especially about the future.
The primary reason China does the world's manufacturing isn't that they are better at it. The primary reason is that it is not a path for richer countries to get even richer. We outsource that activity because we have better uses of labor.
I'd also argue that whatever gains and improvements they make will be things that mostly transfer and can be imitated. One need only look at continuous quality improvement and just in time processes first implemented in the car industry by Japan.
For once we agree. If robotic factories are the future, and labor costs no longer confer a critical advantage, then factories can be located practically anywhere, and you save on transportation costs by reshoring.
> then factories can be located practically anywhere
Theoretically yes, but there are huge gains from industrial agglomeration which China is now reaping. Factories need hundreds or thousands of inputs, and it is a great advantage to have producers or suppliers of those inputs close at hand rather than scattered over a large area or, heaven forfend, abroad, when your logistics depends on the whims of customs officials, pirates, longshoremen unions, and so on.
I wasn't thinking of anything close to what you state but I agree with that too. ... At least to the end extent that manufacturing a good at many sites isn't more of expensive than the transportation from one or very few sites.
The short answer is anything that pays more than the wages in China and other low cost countries, including working at Starbucks or cleaning other people's houses.
No matter how much some people lament the disappearance of low-skilled assembly line manufacturing jobs, they are mostly miserable work. The only way people in the US will take those jobs is if they have no other choice or they pay significantly more. People talk about Henry Ford paying higher wages to workers without understanding why he did that. He HAD TO pay more in order to attract enough workers.
Exceptions:
1 has to be produced somewhere govt does interfere too much with the process or add too much extra costs.
2 not applicable to most things that are custom made, require more skilled labor, or quick turnaround.
3 not applicable to critical items for which using foreign sources could pose some kind of security risk.
I thought the story was Henry paid his people more so they could afford the cars they were manufacturing. At that time, farms in Michigan were collapsing and a factory job was a good gig and there were plenty of workers. So goes the story we were told in Michigan.
Honestly, I don't pay much attention or lend credence to historical narratives developed by special interests intent on pushing political imperatives, nor to those pushing opposite narratives to support personal political beliefs. The world is more complicated than any narrative I've ever read about the auto industry.
Reporting from Wuhan... Robots. I've been in the factories. Robots are building robots. Auto plants are essentially all robotic with human handlers walking around should something go wrong, which doesn't seem to happen. Xiaomi now has a fully automated robotic factory cranking out their cell phones (that are darn nice BTW) at the rate of 60 per minute...one phone every second. No humans involved. Xiaomi also has their auto mfg. nearly all robotized, the cars are beautiful, and are so price competitive there's a year wait to get their new SU7.
Couple all that into supply chain logistics where everything, and I do mean everything, is next door, all feeding into rail spurs going directly to computer automated loading and unloading gantries with related dockage along the Yangtze all the way from Chongqing to Shanghai, with most of it located in the middle to lower reaches in the Yangtze River Valley Development Zone...that's just the Yangtze. The same thing exists down in the Pearl River Zone. It is truly the factory for the world. I could go on, but won't.
America is more than flat footed; it's chopped off at the ankles. There's not going to be catch up. Some gains here and there, but right now, it's looking like a horizon job...i.e., in a sailing race when one boat is over the horizon and the rest of the fleet is tacking around trying to find a favorable breeze.
Sorry to say it, but I can't unsee what I saw.
Thanks for another good note from China.
Are the normal folk in China feeling a lot richer? Most having most material food, clothes, cars, bikes, computers, housing?
Now I’m wondering how many drone robots and drone robot invaders China would be willing to lose to invade Taiwan.
It's not possible to describe "normal folk in China" due to vast socio-economic differences in societal structure between the US and the PRC. "Normal" isn't a useful adjective. The middle class is doing fine and enjoying life. The majority of citizens are not middle class; life is hard. Like really fxxxing hard. There are HUGE disparities and inequities between urban and rural citizens.. If you're rural, which is several hundred million people, you're still shut out from the economic advances. It's a wildly diverse and diffused economic picture, one that is not portrayed in the West in any coherent manner. Direct comparisons between here and the US do not lend themselves to understanding what's going on.
The robot part...is freaky. Unless the American robot makers are keeping everything secret and under wraps so we can't know reality, China is way ahead of the US in robot tech and AI integration. Just one little metric.... Robot taxis aren't something anyone wonders about. They're all over the place in Wuhan and work perfectly.
Waymo is leading the way in the US but it is slowed by regulations meant for humans
Yeah, I know. It’s amazing to ride in one over here. Very smooth, everything works. It’s a very weird feeling.
China’s has a demographic problem just like S Korea and Japan, just a decade or so behind. Then there’s a financial train wreck that is coming for them.
China is also importing a lot of oil and food which could be disrupted if they get feisty about Taiwan.
Maybe, but Western media dwells on what makes China look tenuous or awful. Financial train wreck...they're handling it remarkably well. Demographic time bomb...we'll see. Lots of Econ PhD's talk about this as inescapable doom. Consider the source.
They will get feisty about Taiwan. No if; when. There's a reason Bid Daddy has an "ironclad" partnership with Vlad. Oil, gas. Food? You need to visit China and take a few train rides cross country. Chinese dig in and endure. America is the place that freaks out, not China.
In short, don't read FP, The Economist, or any Western media and imagine you know what's going on in China.
Don’t think much of Peter Zeihan then?
I have been to China a few times working at refineries and was underwhelmed by the operations. The problem I saw was silos and deference to superiors that stifled production and operations.
Peter is a bozo. He's got his community, he presents well, speaks authoritatively, etc., etc., but he weaves together a bunch of decontextualized nonsense that fits his already determined narrative. A few times in China is not enough times. When those times are matters a lot. Anything more than 5 years ago isn't relevant. I can't comment on the silos because I don't know where they're at or when they were developed.
I'm not cheerleading for China. I've been in country for 15 years. I'm just reporting facts. The differences in everything in just the last 5 years are overwhelming. Of course, anything can happen. There's evidence for everything. I can list a few disasters, but I can also "what about" any negative predictions/observations for as long as anyone wants to listen.
And really...if you're seriously listening to Pete and imagining he knows the future...you really need to find other information sources.
Who or what do you recommend, besides yourself 😏?
I don't recommend myself. I occasionally cite specific things that I see with my own eyes. If you want to know, read a LOT of stuff. There's a lot of it in Substack. China Thought Express leans negative on China, but not Peter Zeihan predicting the future type negative. Pekingnology is just plain good info. Nathan Whitaker has his Rural China Substack Blog. Robert Wu is excellent, very fair minded Chinese guy in the Mainland that reports reality both sides. The Monitor often has good stuff. There's a guy that just goes by Bill that I just stumbled upon that spent years in China doing business and he's got some really good insights into how stuff works here. Bill... Do a search for Bill with China stuff. I read Western media too because it's important to see all sides, although most of it is simpleton biased decontextualized nonsense.
The main point I'd make is don't read one thing and imagine it's the single reality. China is a very large and very complicated place and it's real easy to fall into ideological bias. It's not all one thing.
Also...keeping it simple...South China Morning Post. And, seriously, Xinhua, the government mouthpiece. Yes, there's the obvious propagandistic stuff, but it's full of useful news, as in real stuff that's happening now.
Lots of essays in Substack. I was searching and finding more than I feel like linking to.
Great links, especially from Semi-analysis and Klapper. The first because of my career in supply chain/logistics and continuing interest in technology impact there. The second because Klapper's essay had a certain profoundness to it that correctly, I believe, articulates a paradox of knowledge and faith. Acknowledging doubts requires honesty with oneself, and can lead to a more nuanced and mature knowledge - or faith. And as I look around lately, honesty is something we need more of.
Great links! And comments. If China is so leading in robotics, then it’s very very good that Trump put some tariffs on their outputs. And more would have been better. We car makers in Slovakia better push the EU to stop Chinese mfg dominance in robots and everything else, especially cars. Robots making robots will hugely reduce the value of most menial human labor. Why pay low wages to an illegal when one can buy a robot to do it?
AI not being allowed to say “I don’t know” seems the essence of the hallucination problem. Spock’s willingness to say “approximately 4.3 seconds “ was a charming nod to humility yet preciseness. Klapper’s post, combined Zhao on searching thru more possible right answers, reinforces my idea that error detection and correction remains a key problem not yet solved. But the checking of differences from different ai model answers seems likely to avoid the avoidable untruths. Likely helped hugely with a human SME (subject matter expert) to focus fact checking on likely untruths.
Mollick’s great team-AI work implies mostly, maybe exclusively, two person teams. So now I’m wondering about multi-AI teams with just one human. Or maybe two for big creativity.
Underdiscussed is the difference in deciding what goals should be sought vs the best way to achieve the chosen goal. Humans will remain most important in choosing the goal, while ai will better optimize how to achieve it. As ai navigation now does in replacing map reading with clear step by step directions, to which a driver must pay somewhat close attention to. And so many drivers are now very weak at reading and understanding maps.
"At the moment, the West is caught flatfooted"
Balderdash. What was it Berra said? Something like predictions are different fficult, especially about the future.
The primary reason China does the world's manufacturing isn't that they are better at it. The primary reason is that it is not a path for richer countries to get even richer. We outsource that activity because we have better uses of labor.
I'd also argue that whatever gains and improvements they make will be things that mostly transfer and can be imitated. One need only look at continuous quality improvement and just in time processes first implemented in the car industry by Japan.
For once we agree. If robotic factories are the future, and labor costs no longer confer a critical advantage, then factories can be located practically anywhere, and you save on transportation costs by reshoring.
> then factories can be located practically anywhere
Theoretically yes, but there are huge gains from industrial agglomeration which China is now reaping. Factories need hundreds or thousands of inputs, and it is a great advantage to have producers or suppliers of those inputs close at hand rather than scattered over a large area or, heaven forfend, abroad, when your logistics depends on the whims of customs officials, pirates, longshoremen unions, and so on.
I wasn't thinking of anything close to what you state but I agree with that too. ... At least to the end extent that manufacturing a good at many sites isn't more of expensive than the transportation from one or very few sites.
"We outsource that activity because we have better uses of labor." Honest question...
What are those better uses for labor?
The short answer is anything that pays more than the wages in China and other low cost countries, including working at Starbucks or cleaning other people's houses.
No matter how much some people lament the disappearance of low-skilled assembly line manufacturing jobs, they are mostly miserable work. The only way people in the US will take those jobs is if they have no other choice or they pay significantly more. People talk about Henry Ford paying higher wages to workers without understanding why he did that. He HAD TO pay more in order to attract enough workers.
Exceptions:
1 has to be produced somewhere govt does interfere too much with the process or add too much extra costs.
2 not applicable to most things that are custom made, require more skilled labor, or quick turnaround.
3 not applicable to critical items for which using foreign sources could pose some kind of security risk.
I thought the story was Henry paid his people more so they could afford the cars they were manufacturing. At that time, farms in Michigan were collapsing and a factory job was a good gig and there were plenty of workers. So goes the story we were told in Michigan.
Yes, that is the story. Does it makes sense?
Honestly, I don't pay much attention or lend credence to historical narratives developed by special interests intent on pushing political imperatives, nor to those pushing opposite narratives to support personal political beliefs. The world is more complicated than any narrative I've ever read about the auto industry.
Absolutely. But people push them as fact in discussions all the time.