On the morning of July 3rd, it was announced that since Kai-Fu Lee publicly announced the establishment of AI2.0 new company at the end of March this year, after three months of preparation, the company has officially been named “Lingyi Wanwu” in Chinese and its official website (01.AI) has also gone live.
It is reported that 01.AI is committed to promoting a revolution in technology, platform, and application across multiple levels with a breakthrough in pedestal-based large models for AI 2.0.
According to the introduction, 01.AI has chosen the most difficult and challenging path of self-developed models in establishing AI 2.0 large language model. To achieve this, 01.AI is building a powerful experimental platform, constructing thousands of GPU clusters for training, optimization, and inference. The aim is to increase the effective parameter quantity of data and improve the density of high-quality data used in order to build continuous, controllable, and scientific training capabilities.
At present, the technology experts in the seven major directions of 01.AI are from various domestic and foreign technology giants. They have a high proportion of doctoral degrees and have already achieved internal testing with a model scale of tens of billions within three months. They are currently expanding towards a model scale of 300 to 700 billion parameters.
Kai-Fu Lee stated, “01.AI welcomes outstanding talents with AI 2.0 technical capabilities and AGI beliefs to join us in building a brand new platform for AI 2.0 and accelerating the arrival of AGI.” When introducing the 01.AI brand, he pointed out that “01.AI represents the digital world, from zero to one, even encompassing everything in the universe. It symbolizes the ambition of Zero-One Intelligence to empower all things.”