To be fair, the US did have a lead in this tech initially. DeepSeek was actually bootstrapped off the llama model that was initially open sourced by meta because they were basically throwing in the towel when they realized they were far behind openai and google. So, a lot of the initial research did come from the US.
Oh damn that’s news to me. But if I understand correctly, it does make sense because China seems to be more focused on AI’s application to the real world
Oh yeah absolutely, Chinese companies are basically treating LLMs like shared infrastructure, and are focusing on building things on top of it. It’s similar to the way companies use Linux today where nobody is actually trying to monetize it directly, but people build useful platforms on top of it. Another huge difference with China is that it has a massive industry, so there are more niches to apply this tech as a result. You can try doing things like factory automation, robotics, infrastructure monitoring, etc. If you don’t have these types of industries or mass infrastructure, then you can’t really apply these tech to these domains. So, AI research in China is inherently more grounded than it is in the US.
To be fair, the US did have a lead in this tech initially. DeepSeek was actually bootstrapped off the llama model that was initially open sourced by meta because they were basically throwing in the towel when they realized they were far behind openai and google. So, a lot of the initial research did come from the US.
Oh damn that’s news to me. But if I understand correctly, it does make sense because China seems to be more focused on AI’s application to the real world
Oh yeah absolutely, Chinese companies are basically treating LLMs like shared infrastructure, and are focusing on building things on top of it. It’s similar to the way companies use Linux today where nobody is actually trying to monetize it directly, but people build useful platforms on top of it. Another huge difference with China is that it has a massive industry, so there are more niches to apply this tech as a result. You can try doing things like factory automation, robotics, infrastructure monitoring, etc. If you don’t have these types of industries or mass infrastructure, then you can’t really apply these tech to these domains. So, AI research in China is inherently more grounded than it is in the US.