Despite some controversy over the extent to which Yi-34B borrows from Llama 2, the Chinese model functions differently due to the data it has been fed. While China's AI industry is still seen as lagging behind the US, Lee believes that the open-source approach will be crucial for any country to fully leverage AI. He argues that this approach can help reduce inequality and foster innovation among researchers, students, entrepreneurs, and hobbyists.
Key takeaways:
- 01.AI, a Beijing-based startup, has released an open-source AI model called Yi-34B that outperforms Meta's Llama 2 and is gaining recognition in the West. The company has also launched a multimodal AI model, Yi-VL-34B, that can process and discuss images.
- The startup, founded by Kai-Fu Lee, has raised $200 million in investment from Alibaba and others and is valued at over $1 billion. Lee has a long history in AI, having founded Microsoft’s Beijing lab and led Google’s Chinese business.
- 01.AI is working on developing "AI-first" apps for office productivity, creativity, and social media, aiming to lead the next phase of the AI revolution. The company's open-source language model has already won admirers in the West.
- Despite the success of 01.AI's models, experts suggest that China's AI industry as a whole still lags behind the US. However, Lee argues that an open-source approach is crucial for any country to fully leverage AI.