>> Most major technological advancements are deflationary when widely adopted, because more can be done with less. Generative AI is no different.
Uhhh... no?
You're confusing micro and macro.
Microeconomically, YES, most manufactured products' prices decline over time because of economies of scale and improving technology. The classic example is the TV - a massive flat-panel screen has never been cheaper to the average family.
Macroeconomically, NO. The slow march towards cheap and ubiquitous flat-screen TVs didn't cause any kind of deflation over the last 30 years. Massive technological advances don't cause monetary deflation, they just disrupt industries. Sometimes those disruptions result in inflation, sometimes deflation, but that's a coordination-and-macroeconomic-policy problem, not a technology problem. The steam engine may have touched off massive deflationary spirals, but that was because they didn't have central banking -- in the modern era of central banking, literally no discovery (to wit, the internet, smartphones, memory foam, microchips, etc.) has triggered any similar kind of deflationary spiral. The problem was VERY obviously that we didn't have central banking, not because of technological advances.
I think the argument depends on how large the impact of Generative AI is. I’m talking about Generative AI, not AI itself. I started thinking if Generative AI is really smart. Artificial Intelligence assumes that human beings are intelligent and we have to create a silicon copy of our brains, but are our brains really smart? And, Generative AI is, if my interpretation is correct, trying to create more human beings like artificial intelligence so that we can communicate and recognize their intelligence. Is it really “intelligence” we are trying to achieve? Just like autonomous driving. It is difficult to implement because human drivers are unpredictable (and stupid) so that autonomous driving solutions have to be smart enough to understand our stupidity and illogicality. Is this a human advancement?
Anyway, I don’t deny the impact of Generative AI on productivity. It will be massive.
It’s true that China doesn’t necessarily value AIGC as much as the US. It’s not only a matter of will but also a matter of ability and preferences. At the most abstract level, AIGC is but a higher-order expression of a creative, talkative culture, which China/Chinese people are not good at. Talking, making speeches, writing fancy stories are not as important to China as in the US. However, it’s true that AIGC can only achieve so much, there is still a whole bunch of other technologies to pursue. My own analogy is that: you won’t need a literature PhD to drive a truck. (Btw, the person fired is not head of NPPA though, who is a minister-level person. it’s only the head of the bureau in the 中宣部 that’s responsible for liaisoning with and managing NPPA. NPPA reports to 中宣部 now)
No Reuters was not wrong: “Feng Shixin was removed last week from his position as head of the publishing unit of the Communist Party's Publicity Department, the sources said. The department oversees the National Press and Publication Administration (NPPA) which in turn regulates China's vast video games sector.”
China can definately not afford to be left behind in Gen AI and open-source gives them an easy path to catching up. The foreign investing drop is poor timing for their efforts however. A lot your points well, I disagree with them. Baidu's Ernie his 100 million likely faster than ChatGPT did in less time.
A majority of the best AI researchers are Chinese, even if educated in America. China will dominant several industries in the future and most likely AI companies as well. The rise of Temu and TikTok Social commerce in 2024 will shock a lot of people in E-commerce, already will become dominant in EV sales globally.
It's easy to point out China's struggles and serious economic depression and issues. But Generative AI is actually a good path forward for stimulating the economy and job market again, just like Alibaba trying to pivot in it.
China's ceiling in how it implements Generative AI and LLMs and SLMs is much higher than in the West to be honest. More data, that they turned on their BigTech might be a boon in the long run to simulate innovation. You think the Mag seven care about real innovation?
Turning Generative AI advantages in the military is something China is definately a world leader in doing, I wouldn't trust DARPA or even the splurge in Defense startups in the U.S. to be even on par. A lot of the best AI papers I read are from China or Hong Kong or Singapore.
Just like the bloated 2023 stock market in the U.S., Generative AI is a huge lever China can and will pull in 2024 or maybe in 2025. To think that they cannot or won't isn't even barely rational. The stakes are too high Kevin.
Great article Kevin, especially tying it back to soft tech vs hard tech. That's a distinction the government has emphasized significantly over the last three years and the CCP's tolerance for "soft tech" has declined significantly since Jack Ma's infamous remarks in Nov 2020.
It's difficult to understand the CCP's AI strategy going forward. Given their aversion to GenAI, I believe they thought AI would be a tool that would drive innovation in "hard sectors", such as manufacturing, pharma, defense, robotics. After ChatGPT, the world finally understands the 'intelligence' in AI. Without an investment in intelligence or GenAI, it's not clear how one will benefit from the applications of AI.
Their cautious approach will backfire long-term. On a side-note, I wonder if there is a cultural reason why China, EU, and Japan are all champions of "hard tech" over "soft-tech"/software. The lack of competitiveness with the US is one thing, but the underlying sentiment remains prevalent in all of these countries.
You're confusing what the CCP wants with what Chinese people want. When ChatGPT was first released, there was a palpable sense of excitement on Chinese social media and many people with connections overseas were able to sign up and try it. By denying China generative AI, the CCP is only making it harder for the Chinese economy, which is increasingly services based and productivity driven, to compete with the rest of the world. This is just another example of why the command economy doesn't work.
>> Most major technological advancements are deflationary when widely adopted, because more can be done with less. Generative AI is no different.
Uhhh... no?
You're confusing micro and macro.
Microeconomically, YES, most manufactured products' prices decline over time because of economies of scale and improving technology. The classic example is the TV - a massive flat-panel screen has never been cheaper to the average family.
Macroeconomically, NO. The slow march towards cheap and ubiquitous flat-screen TVs didn't cause any kind of deflation over the last 30 years. Massive technological advances don't cause monetary deflation, they just disrupt industries. Sometimes those disruptions result in inflation, sometimes deflation, but that's a coordination-and-macroeconomic-policy problem, not a technology problem. The steam engine may have touched off massive deflationary spirals, but that was because they didn't have central banking -- in the modern era of central banking, literally no discovery (to wit, the internet, smartphones, memory foam, microchips, etc.) has triggered any similar kind of deflationary spiral. The problem was VERY obviously that we didn't have central banking, not because of technological advances.
I think the argument depends on how large the impact of Generative AI is. I’m talking about Generative AI, not AI itself. I started thinking if Generative AI is really smart. Artificial Intelligence assumes that human beings are intelligent and we have to create a silicon copy of our brains, but are our brains really smart? And, Generative AI is, if my interpretation is correct, trying to create more human beings like artificial intelligence so that we can communicate and recognize their intelligence. Is it really “intelligence” we are trying to achieve? Just like autonomous driving. It is difficult to implement because human drivers are unpredictable (and stupid) so that autonomous driving solutions have to be smart enough to understand our stupidity and illogicality. Is this a human advancement?
Anyway, I don’t deny the impact of Generative AI on productivity. It will be massive.
It’s true that China doesn’t necessarily value AIGC as much as the US. It’s not only a matter of will but also a matter of ability and preferences. At the most abstract level, AIGC is but a higher-order expression of a creative, talkative culture, which China/Chinese people are not good at. Talking, making speeches, writing fancy stories are not as important to China as in the US. However, it’s true that AIGC can only achieve so much, there is still a whole bunch of other technologies to pursue. My own analogy is that: you won’t need a literature PhD to drive a truck. (Btw, the person fired is not head of NPPA though, who is a minister-level person. it’s only the head of the bureau in the 中宣部 that’s responsible for liaisoning with and managing NPPA. NPPA reports to 中宣部 now)
good to know re: firing. Looks like Reuters mis-reported?
No Reuters was not wrong: “Feng Shixin was removed last week from his position as head of the publishing unit of the Communist Party's Publicity Department, the sources said. The department oversees the National Press and Publication Administration (NPPA) which in turn regulates China's vast video games sector.”
China can definately not afford to be left behind in Gen AI and open-source gives them an easy path to catching up. The foreign investing drop is poor timing for their efforts however. A lot your points well, I disagree with them. Baidu's Ernie his 100 million likely faster than ChatGPT did in less time.
A majority of the best AI researchers are Chinese, even if educated in America. China will dominant several industries in the future and most likely AI companies as well. The rise of Temu and TikTok Social commerce in 2024 will shock a lot of people in E-commerce, already will become dominant in EV sales globally.
It's easy to point out China's struggles and serious economic depression and issues. But Generative AI is actually a good path forward for stimulating the economy and job market again, just like Alibaba trying to pivot in it.
China's ceiling in how it implements Generative AI and LLMs and SLMs is much higher than in the West to be honest. More data, that they turned on their BigTech might be a boon in the long run to simulate innovation. You think the Mag seven care about real innovation?
Turning Generative AI advantages in the military is something China is definately a world leader in doing, I wouldn't trust DARPA or even the splurge in Defense startups in the U.S. to be even on par. A lot of the best AI papers I read are from China or Hong Kong or Singapore.
Just like the bloated 2023 stock market in the U.S., Generative AI is a huge lever China can and will pull in 2024 or maybe in 2025. To think that they cannot or won't isn't even barely rational. The stakes are too high Kevin.
Great article Kevin, especially tying it back to soft tech vs hard tech. That's a distinction the government has emphasized significantly over the last three years and the CCP's tolerance for "soft tech" has declined significantly since Jack Ma's infamous remarks in Nov 2020.
It's difficult to understand the CCP's AI strategy going forward. Given their aversion to GenAI, I believe they thought AI would be a tool that would drive innovation in "hard sectors", such as manufacturing, pharma, defense, robotics. After ChatGPT, the world finally understands the 'intelligence' in AI. Without an investment in intelligence or GenAI, it's not clear how one will benefit from the applications of AI.
Their cautious approach will backfire long-term. On a side-note, I wonder if there is a cultural reason why China, EU, and Japan are all champions of "hard tech" over "soft-tech"/software. The lack of competitiveness with the US is one thing, but the underlying sentiment remains prevalent in all of these countries.
You're confusing what the CCP wants with what Chinese people want. When ChatGPT was first released, there was a palpable sense of excitement on Chinese social media and many people with connections overseas were able to sign up and try it. By denying China generative AI, the CCP is only making it harder for the Chinese economy, which is increasingly services based and productivity driven, to compete with the rest of the world. This is just another example of why the command economy doesn't work.