I’ve always found your GPT diffusion argument to be compelling. Yet others, while not disagreeing entirely with your analysis, suggest that the speed of AI development and diffusion is fundamentally different than, say, electricity. Which took much longer (though I can argue that AI is not developing as fast as often portrayed).
Hence there is still a major competitive structural advantage to being a LS-like first mover. Something along the lines of, getting there first is critical, and will eventually help increase the advantages inherent in broad and deep diffusion.
Good critique! Here's my response from p. 4 of the article, in which I argue that in some cases being the LS-like first mover can result in an advantage in GPT diffusion, but it is not determinative (and we often mistakenly treat it as the sole, determinative factor): "It is important to note that some parts of the GPT and LS mechanisms can co-exist. A state’s capacity to pioneer new technologies can correlate with its capacity to absorb and diffuse GPTs. Countries home to cutting-edge R&D infrastructure might also be fertile ground for education systems that widen the pool of GPT-linked engineering skills. However, these aspects of the LS mechanism are not necessary to the GPT mechanism. A state can capitalize on GPTs to become the most powerful economy without monopolizing LS innovation. Moreover, there is a direct conflict between the mechanisms’ expectations regarding impact timeframe and breadth of growth."
I think it's also important to note that in GPTs, there's no single country that is going corner all major innovations in this field; if we're looking at a marathon-type technological competition, being the first-mover by one year might not matter as much in terms of being the leader GPT diffusion. See also the second industrial revolution case evidence, which shows that the U.S. was by no means the innovation leader in the new sectors, but it was still able to become the leader in GPT diffusion.
生日快乐!
I’ve always found your GPT diffusion argument to be compelling. Yet others, while not disagreeing entirely with your analysis, suggest that the speed of AI development and diffusion is fundamentally different than, say, electricity. Which took much longer (though I can argue that AI is not developing as fast as often portrayed).
Hence there is still a major competitive structural advantage to being a LS-like first mover. Something along the lines of, getting there first is critical, and will eventually help increase the advantages inherent in broad and deep diffusion.
How do you answer that critique?
Good critique! Here's my response from p. 4 of the article, in which I argue that in some cases being the LS-like first mover can result in an advantage in GPT diffusion, but it is not determinative (and we often mistakenly treat it as the sole, determinative factor): "It is important to note that some parts of the GPT and LS mechanisms can co-exist. A state’s capacity to pioneer new technologies can correlate with its capacity to absorb and diffuse GPTs. Countries home to cutting-edge R&D infrastructure might also be fertile ground for education systems that widen the pool of GPT-linked engineering skills. However, these aspects of the LS mechanism are not necessary to the GPT mechanism. A state can capitalize on GPTs to become the most powerful economy without monopolizing LS innovation. Moreover, there is a direct conflict between the mechanisms’ expectations regarding impact timeframe and breadth of growth."
I think it's also important to note that in GPTs, there's no single country that is going corner all major innovations in this field; if we're looking at a marathon-type technological competition, being the first-mover by one year might not matter as much in terms of being the leader GPT diffusion. See also the second industrial revolution case evidence, which shows that the U.S. was by no means the innovation leader in the new sectors, but it was still able to become the leader in GPT diffusion.