A very interesting article. In the context of whether China has the capability for genuine innovation, it should be noted that western Biopharma companies have been beating a path to Chinese companies (in reality a stampede) to uncover new drugs of late. In 2024 such in-licensing of assets (from China) constituted around one third of total in-licensed drugs (which is significant). To emphasize, these are novel drugs and not generics or copies of western drugs. Thus, I would say that China is unquestionably a source of innovation.
One aspect of the DeepSeek story I've not yet seen reported is DeepSeek as an act of national philanthropy, basically Liang Wenfeng giving back to the nation after the crackdown on the antisocial success of his trading business.
If this is in fact part of his motivation and thinking, despite his organization's not being a typical state champion and apparently not adopting typical organizational practices, there is an interesting nationalist dimension here.
Great read! The concept of deepseek as an outlier and even unknown or backed in China is fascinating. It eliminates the tendency to lump all Chinese AI firms into a common group an ignorant tendency due to US lack of experience with China. Chinese firms truly are as different as Meta is to OpenAi in the US.
I agree with the general ideas here, but I’m surprised you didn’t mention that DeepSeek used OpenAI’s algorithms to help train their model. It’s looking like this was at least partially a case of technology transfer, and that’s how they avoided some of the huge costs of training their own model.
"This is where DeepSeek diverges from the traditional technology transfer model that has long defined China's tech sector. Instead of relying on foreign-trained experts or international R&D networks, DeepSeek exclusively uses local talent."
This part reminds me of Zhang Yiming—he's very much like Liang Wenfeng, never trained at a foreign university and struggling initially to raise funds due to his grassrootness and seemingly provincial ("土气") background. What's striking is that neither displays the hubris I've consistently encountered among Bay Area founders. This new generation of Chinese tech entrepreneurs actually disenchanted with the American playbook quite early on when silicon valley was still quite self-indulgent.
Организационная и производственная модель DeepSeek это заметное технологическое решение. Но я бы не советовал Китаю отменять или массово перестраивать апробированную модель управления предприятиями и экономикой в целом. Государственное регулирование, инвестиции, планирование ресурсов и kpi - это залог длительного устойчивого развития. Китай 🇨🇳 способен к самостоятельным инновациям с сохранением коммунистического режима в экономике.
В теле есть мышцы синергисты и антагонисты, всё зависит от момента действия, геометрического и математического положения тела во времени и пространстве.
"As development economists would remind us, all technology must first be transferred to and absorbed by latecomers; only then can they innovate and create breakthroughs of their own. Thus, tech transfer and indigenous innovation are not mutually exclusive — they’re part of the same sequential progression. " - This begs the question then whether any country can actually develop tech on its own without theft (as the US, China, and Japan have done) or the stronger power willingly sharing.
Very nice article on Chinese work culture. But I would venture to speculate that this pedagogal education and regimented work ethic is more of a feature than a hinderance with respect to the needs of the CCP. I would think that the CCP finds lateral thinking to be a threat to their hold on authority. It would be interesting to know the flow of the conversation between the government representatives and Wenfeng. Are they impressed or is he on a tight leash? After all, given that they were free to experiment to create the V3 and R1 model, they could have easily fine-tuned out the most obvious anti-CCP rhetoric. But rather they decided on system level guard rails to handle sensitive topics. Moreover, as it is open sourced for download, end users who host the model can now have all this information at their finger tips.
A very interesting article. In the context of whether China has the capability for genuine innovation, it should be noted that western Biopharma companies have been beating a path to Chinese companies (in reality a stampede) to uncover new drugs of late. In 2024 such in-licensing of assets (from China) constituted around one third of total in-licensed drugs (which is significant). To emphasize, these are novel drugs and not generics or copies of western drugs. Thus, I would say that China is unquestionably a source of innovation.
Great great article.
One aspect of the DeepSeek story I've not yet seen reported is DeepSeek as an act of national philanthropy, basically Liang Wenfeng giving back to the nation after the crackdown on the antisocial success of his trading business.
If this is in fact part of his motivation and thinking, despite his organization's not being a typical state champion and apparently not adopting typical organizational practices, there is an interesting nationalist dimension here.
Great read! The concept of deepseek as an outlier and even unknown or backed in China is fascinating. It eliminates the tendency to lump all Chinese AI firms into a common group an ignorant tendency due to US lack of experience with China. Chinese firms truly are as different as Meta is to OpenAi in the US.
Very nice article !
Nice article Jordan!
Super interesting read!
I agree with the general ideas here, but I’m surprised you didn’t mention that DeepSeek used OpenAI’s algorithms to help train their model. It’s looking like this was at least partially a case of technology transfer, and that’s how they avoided some of the huge costs of training their own model.
Great article!
"This is where DeepSeek diverges from the traditional technology transfer model that has long defined China's tech sector. Instead of relying on foreign-trained experts or international R&D networks, DeepSeek exclusively uses local talent."
This part reminds me of Zhang Yiming—he's very much like Liang Wenfeng, never trained at a foreign university and struggling initially to raise funds due to his grassrootness and seemingly provincial ("土气") background. What's striking is that neither displays the hubris I've consistently encountered among Bay Area founders. This new generation of Chinese tech entrepreneurs actually disenchanted with the American playbook quite early on when silicon valley was still quite self-indulgent.
Excellent piece
Организационная и производственная модель DeepSeek это заметное технологическое решение. Но я бы не советовал Китаю отменять или массово перестраивать апробированную модель управления предприятиями и экономикой в целом. Государственное регулирование, инвестиции, планирование ресурсов и kpi - это залог длительного устойчивого развития. Китай 🇨🇳 способен к самостоятельным инновациям с сохранением коммунистического режима в экономике.
В теле есть мышцы синергисты и антагонисты, всё зависит от момента действия, геометрического и математического положения тела во времени и пространстве.
They will develop their own tech, more advanced than that of the US, and then restrict it to them in revenge.
Very insightful and helpful, thanks.
"As development economists would remind us, all technology must first be transferred to and absorbed by latecomers; only then can they innovate and create breakthroughs of their own. Thus, tech transfer and indigenous innovation are not mutually exclusive — they’re part of the same sequential progression. " - This begs the question then whether any country can actually develop tech on its own without theft (as the US, China, and Japan have done) or the stronger power willingly sharing.
Very nice article on Chinese work culture. But I would venture to speculate that this pedagogal education and regimented work ethic is more of a feature than a hinderance with respect to the needs of the CCP. I would think that the CCP finds lateral thinking to be a threat to their hold on authority. It would be interesting to know the flow of the conversation between the government representatives and Wenfeng. Are they impressed or is he on a tight leash? After all, given that they were free to experiment to create the V3 and R1 model, they could have easily fine-tuned out the most obvious anti-CCP rhetoric. But rather they decided on system level guard rails to handle sensitive topics. Moreover, as it is open sourced for download, end users who host the model can now have all this information at their finger tips.
we'll just have to see!