China Resets AI Race(wsj.com)
wsj.com
China Resets AI Race
https://www.wsj.com/tech/ai/chinese-ai-anthropic-mythos-cybersecurity-574b02c2
5 comments
https://archive.is/m1qrN
No the AI race hasn’t been reset. Leading American companies maintain 6-12 month advantages and China releasing these models just raises the floor for everyone and provides no real advantage to China from a security perspective but American companies maintain an important lead.
One could make an argument that they are attempting to “flood the market” with lower priced models to attack American companies that have been investing so much in data center and other buildout but it’s a bad strategy. If it was a good strategy then American companies could just let China get ahead and then do the same copy/paste strategy back to them. But American companies don’t do that because the 6-12 month lead is a meaningful and worthwhile advantage.
I don’t intend to disparage or discount China or the hard work Chinese engineers (why don’t they let in more immigrants? Wouldn’t their models be better by harvesting the world’s talent? Interesting. Anyway.) put into making these advancements but I find the general analysis, pearl-clutching, and fear mongering to be a bit annoying.
Think about it this way. If these open-weight models and such were the threat that others say they are, Americans could just stop training models, wait and let China spend all the money, and then release their own open weight models and “reset the AI race”. Why do we spend money when we don’t have to? Well it’s because the premise is bullshit and it does matter that we have a lead and it does matter that we continue to train models as we are doing.
One could make an argument that they are attempting to “flood the market” with lower priced models to attack American companies that have been investing so much in data center and other buildout but it’s a bad strategy. If it was a good strategy then American companies could just let China get ahead and then do the same copy/paste strategy back to them. But American companies don’t do that because the 6-12 month lead is a meaningful and worthwhile advantage.
I don’t intend to disparage or discount China or the hard work Chinese engineers (why don’t they let in more immigrants? Wouldn’t their models be better by harvesting the world’s talent? Interesting. Anyway.) put into making these advancements but I find the general analysis, pearl-clutching, and fear mongering to be a bit annoying.
Think about it this way. If these open-weight models and such were the threat that others say they are, Americans could just stop training models, wait and let China spend all the money, and then release their own open weight models and “reset the AI race”. Why do we spend money when we don’t have to? Well it’s because the premise is bullshit and it does matter that we have a lead and it does matter that we continue to train models as we are doing.
Interesting take. One question though: is it even possible for US companies? I am not so sure. Investors put a lot of money in those companies and they expect return. Doing nothing for 6-12 months waiting for China would probably make a lot of important people very unhappy.
I guess the central point here is that many view this as a race and one in which if you are ahead the competitor can undercut you and being in the lead doesn't matter and instead cost matters - China's model can do XYZ at X% the cost of Claude!! So doomed! But if that's true for the United States it's of course true for China too. In other words, if American companies stopped investing they would have to stop investing too otherwise they would wind up in the same situation, it's like a game of hot potato. I don't buy it because it doesn't make sense. I think the 6-12 month lead time for American companies actually matters and China doesn't have a great hand to play here. At best they can catch up, if they pull ahead then American companies can just play the same hand that China seems to be playing.
They're trying to (and other commentators are as well) look at this like a manufacturing play where you flood the market with low-cost goods but I think that's not just simplistic but wrong too.
They're trying to (and other commentators are as well) look at this like a manufacturing play where you flood the market with low-cost goods but I think that's not just simplistic but wrong too.
What is wrong too?
The plan to create a lot of cheap models? That it is what is happening but it is bad plan?
Or the opinion that it is what is happening but it is wrong opinion because they do not plan to do that but i.e. China is just bigger with more people so just more companies are releasing more models?
Maybe both is happening a lot of choices will make more people go with Chinese models, which means less money to the US which means more problem to US companies. But at the same time they are cheaper and less of quality because this is China, this is what they does.
Personally I think this much more complicated as everything in life.
Maybe both is happening a lot of choices will make more people go with Chinese models, which means less money to the US which means more problem to US companies. But at the same time they are cheaper and less of quality because this is China, this is what they does.
Personally I think this much more complicated as everything in life.