Kunpeng Project
Zhou Hongyi: The AI gap between China and the United States is about 1 to 2 years, and the direction of innovation is the key.
AI has been a hot topic in the tech world for more than a year, with an AI war unfolding around AI from Chat GPT to SORA.
Many people say that from Chat GPT to SORA, you can see the huge difference between Chinese and American artificial intelligence, because Chat GPT and SORA are both American and American, so we can only keep up.
So, how different is AI between China and the United States? Recently, there was an exclusive interview with Zhou Hongyi, and he gave a more reasonable answer to this question.
He said that there is no technical gap in AI lithography machines and chips, and whether it is Chat GPT or SORA, they are all software-based, and it should be able to make up for it in a year or two at most.
He said that as soon as OpenAI releases a new technological development trend, then in China, its followers will flock to it, that is, in a year or two, China will be able to catch up.
But Zhou Hongyi also said that if others open up resources, their technology will also develop. In fact, this means that China's original technology is still relatively weak, that is, in this field, we are much worse than the United States.
What does this phrase mean? In fact, it's a big question, and it's a big one.
The United States can discover the original direction of development, carry out directional innovation, and find a way to start from scratch, such as the emergence of Chat GPT and SORA, which are all processes from scratch.
We are weak in this field, and it is very difficult to start from scratch, so we can't create chargpt or sora.
But now, with the launch of ChatGPT and Sora, it's like watching other people start from nothing, and then we go and learn and quickly go from 0 to 1.
So, with ChatGPT, we immediately did a large number of large-scale modeling, and the speed from one to ten, far more than other companies, take the large model as an example, the domestic large model is almost a monopoly, and GPT4 has been squeezed into the top ten.
What are your thoughts on this issue? Do you agree with Zhou Hongyi's statement? Do you think that artificial intelligence between China and the United States will only take a year or two, or a few years? In fact, in my opinion, in fact, many years have passed, and there is still one to two years to go from zero to one because of the need to learn from others.
But in the absence of examples, it is impossible for him to start from scratch, let alone from one to ten, and the most critical problem is technological innovation, and this is not something that can be solved in a short time.