This Chinese Startup Is Winning the Open Source AI Race

AI, a startup based in Beijing, has made waves in the field of artificial intelligence by releasing its own open source language model that surpasses the capabilities of models like Llama 2. The startup’s model, called Yi-34B, quickly rose to the top of rankings maintained by Hugging Face, a startup that compares AI language models across various benchmarks. Modified versions of Yi-34B continue to perform well on these rankings and other leaderboards.

Unlike OpenAI and Google, which tightly control their AI technology, 01.AI is giving away its models in the hopes of building a dedicated developer community that can contribute to the creation of innovative AI applications. The company, founded in 2023, has already raised $200 million in investment and is valued at over $1 billion. Its founder and CEO, Kai-Fu Lee, is a prominent investor and AI researcher who previously worked at Microsoft and Google.

Lee envisions AI leading the next phase of the AI revolution by developing “killer apps” based on language models. The company aims to create productivity tools and social media platforms that are more intuitive and user-friendly. While 01.AI’s apps have not been launched yet, their open source language model has gained recognition in the West as being highly capable, even compared to models with larger parameter sizes.

Lee’s career in AI has been notable, starting from his early research on speech recognition systems. He worked at Apple, Silicon Graphics, and Microsoft before becoming a champion of the Chinese AI industry. He has advocated for collaboration between the US and China in the field of AI, and his book “AI Superpowers” discusses the potential of Chinese AI labs and companies to rival those in the US.

Overall, AI is positioning itself as a leading contender in the AI race, aiming to develop groundbreaking applications that leverage the power of language models and cater to global audiences.