Transcription
I think China's already won the tech race with the US and left Europe far behind. Um, and I think frankly, in Europe, we've been mostly asleep for the last 20 years in this regard. Um, a lot of people disagree with me. So let me just cite a couple of pieces of evidence.
Uh, if you look on the website of the ASPI, that's a think tank in Australia, they have something called the uh, the China Tech Tracker. Um, and it says at the moment that China is leading the US in 66 out of 74 advanced technologies, and the US is only ahead in eight. Uh, if you look at the latest Nature Index uh of uh uh cited u sorry, elite papers that have been cited, elite scientific papers that have been cited, uh, China is far ahead of every other country in the world. Uh, if you look at the number of patent applications that the World Intellectual Property Office of the United Nations, China has 1.8 million applications pending, and the US only has 500,000.
So, um, I think we can safely say, um, whatever your perspective on this view is, that China is at least a peer competitor of the United States, and I would say has already won. I would say, you know, the US and China's kind of neck and neck on AI at the moment. I mean, Jensen Huang said uh recently that he thought China was ahead, and then he retracted that the next day. He said China was not quite ahead. So, I mean, clearly there's a race going on, and this is very much in the mind of Chinese and and of Americans.
But there are two big differences uh in the way the race is being run. The first is that most of the Chinese uh AI models that are doing so well, So Kimmy K2, the Q1 suite from Alibaba, um, and, uh, and some other ones as well, they're all open source, right? Whereas the leading American, uh, AI models are not open-source. The other crucial difference is that if you, the price of operating the Chinese models is a fraction of the price of operating the US models. So we're talking about one-tenth or less in terms of the million outbook tokens uh that that you use to judge the cost of operating an AI model.
So uh this is why uh if in the case of Alibaba Q1, they've already got a billion customers worldwide, and that leaves the American AI models in the dirt. So uh there's a lot of interesting stuff going on there. I'm not going to call it. Um, you know, the Americans call it [laughter] too late. >> The Americans, if the Americans could come up with an incredible moonshot, that's what Americans do, and they could win the whole thing. Or there could be a quantum, you know, breakthrough that changes the whole model. I don't, I'm not going to call it, but it is a fascinating thing to watch.