Alibaba’s AI Ambitions
Alibaba, the Chinese e-commerce behemoth, has made a significant stride in the realm of artificial intelligence. The company announced its decision to open its artificial intelligence model to third-party developers. This strategic move is aimed at amplifying the usage of its product and positioning Alibaba as a technological frontrunner.
The implications of this decision are vast. Not only does it place Alibaba in direct competition with U.S. tech titan Meta, but it also presents a potential challenge to OpenAI, the organization behind the widely recognized AI chatbot, ChatGPT.
Earlier in April, Alibaba unveiled its large language model (LLM) named Tongyi Qianwen. For those unfamiliar, an LLM is an AI model that undergoes training on vast data sets. This forms the foundation for generative AI applications, like ChatGPT, known for producing human-esque responses to user queries.
Tongyi Qianwen boasts capabilities in both English and Chinese content generation. It comes in various model sizes, with some having as many as seven billion parameters, a metric indicating its computational power.
In a move that underscores Alibaba’s commitment to the global AI community, the company will open-source its seven-billion-parameter model, Qwen-7B. Additionally, a version tailored for conversational applications, Qwen-7B-Chat, will also be made available. This initiative allows researchers, academics, and corporations worldwide to harness the model for their generative AI applications. They can do so without the need to train their systems, translating to significant time and cost savings.
However, there’s a catch. Companies boasting over 100 million monthly active users will need a royalty-free license from Alibaba.
While this open-source approach might not directly translate to licensing revenue for Alibaba, it’s a strategic play. The widespread distribution of its technology will likely drive more users to its AI model.
This development comes at a pivotal moment for Alibaba. The company is aggressively pushing its cloud computing division, with AI investments at the forefront. They view cloud computing as a pivotal growth area, promising substantial profitability.
Meta, earlier this year, also took the open-source route with its Llama large language model. Collaborations with other tech giants are in the pipeline to enhance its adoption. Microsoft, for instance, announced its decision to make Meta’s Llama 2 accessible on its Azure cloud-computing service.
While Alibaba hasn’t publicized any such collaborations, the success of its LLM in the market might make it an enticing proposition for cloud providers.
To showcase the capabilities of Tongyi Qianwen, Alibaba has already developed applications. A notable mention is Tongyi Wanxiang, an AI tool capable of generating images based on prompts.
