Some good points in this piece, but I think it conflates diffusion of AI throughout the economy (which is desirable for growth) and integrating AI into physical production processes (which China's AI education is focusing on). As the Macro Polo article (https://macropolo.org/chinese-university-ai-ambitions/) cited in the piece notes, China faces challenges integrating AI into industrial processes (e.g. due to risk aversion among Chinese enterprises) and there is a mismatch between AI education and industry demand for AI.
Services currently account for 55% of Chinese GDP and experts say it needs to increase that percentage further to reduce problems caused by excess manufacturing capacity, so it doesn't seem sensible for applied AI courses to be so weighted towards physical production processes.
This is a solid point, there is definitely hesitation by traditional manufacturing industries towards adopting AI. However, the Applied AI majors serve a duel role, lowering the cost for companies to adopt ML in their factories by tackling skilled labor supply, while also diffusing it to “rust belt” regions. Lack of AI adoption is a multifaceted problem across countries but China is using these programs as a policy lever to concretely tackle the labor issue and I think it is the clearest, most effective first step for the industrial sector.
Regarding the service economy point, like I mentioned in the piece, Chinese policy makers are very consciously choosing to lean into high-skilled manufacturing like Germany as opposed to going down the American route of having a service based economy. This is a tougher road to head down but Germany proves it is doable.
The U.S. approach is not distributed it is totally elitist. AI for the 1%. China can use or steal the U.S. leading edge research and focus on applications. Smart. And will help millions not just an elite group in Silicon Valley.
If we can pause the hyperventilating about AGI long enough to focus on what it takes to put AI into production, we might begin to see those promised productivity improvements across the U.S.
I can always tell who’s tried to put AI into production at scale and who looks at it mostly theoretically: how much they focus on the hard work involved in building the AI engine or data pipelines. Which includes, as this post describes so well, a critical (and probably different) education and training component.
Some good points in this piece, but I think it conflates diffusion of AI throughout the economy (which is desirable for growth) and integrating AI into physical production processes (which China's AI education is focusing on). As the Macro Polo article (https://macropolo.org/chinese-university-ai-ambitions/) cited in the piece notes, China faces challenges integrating AI into industrial processes (e.g. due to risk aversion among Chinese enterprises) and there is a mismatch between AI education and industry demand for AI.
Services currently account for 55% of Chinese GDP and experts say it needs to increase that percentage further to reduce problems caused by excess manufacturing capacity, so it doesn't seem sensible for applied AI courses to be so weighted towards physical production processes.
This is a solid point, there is definitely hesitation by traditional manufacturing industries towards adopting AI. However, the Applied AI majors serve a duel role, lowering the cost for companies to adopt ML in their factories by tackling skilled labor supply, while also diffusing it to “rust belt” regions. Lack of AI adoption is a multifaceted problem across countries but China is using these programs as a policy lever to concretely tackle the labor issue and I think it is the clearest, most effective first step for the industrial sector.
Regarding the service economy point, like I mentioned in the piece, Chinese policy makers are very consciously choosing to lean into high-skilled manufacturing like Germany as opposed to going down the American route of having a service based economy. This is a tougher road to head down but Germany proves it is doable.
The U.S. approach is not distributed it is totally elitist. AI for the 1%. China can use or steal the U.S. leading edge research and focus on applications. Smart. And will help millions not just an elite group in Silicon Valley.
You nailed it Steve. The Truth. As everything else the elites will profit from it, that’s probably about it as usual. 👍 thx
Diffusion, diffusion, diffusion.
If we can pause the hyperventilating about AGI long enough to focus on what it takes to put AI into production, we might begin to see those promised productivity improvements across the U.S.
I can always tell who’s tried to put AI into production at scale and who looks at it mostly theoretically: how much they focus on the hard work involved in building the AI engine or data pipelines. Which includes, as this post describes so well, a critical (and probably different) education and training component.