Big factories roll out models, small factories roll out applications, how do ordinary people cope with the AI wave?
ChatGPT exploded in popularity in March this year, there has been a trend of "preparing for war" in large-scale models in China. Major manufacturers have made heavy bets: Baidu, which has been in AI for many years, launched "Wen Xin Yiyan" and vowed to fight ChatGPT. Huawei "Pangu" has swept the industry, and Ali's "Tongyi Qianwen" has landed in the office field... The popularity of large models seems to have shed a ray of light into the long-standing gloom that "big factory traffic has peaked."
Not long ago, on August 31, "Wen Xin Yi Yan" shot to the top of the App Store free list HE Tuber after it was opened to the public. Baidu official showed that Wen Xin Yi Yan's users exceeded the 1 million mark on the first day of its launch.
Prior to this, many large model products including Wen Xinyiyan required internal testing applications. On March 16 this year, Wenxinyiyan officially launched internal testing. The first batch of users can log in through the invitation code and experience the product on Wenxinyiyan’s official website. In early July, Baidu launched the Wen Xin Yi Yan Apple App, but users still needed to qualify for internal testing to experience the corresponding functions.
This time, Baidu AI’s large model reached the top in one fell swoop, and it seemed to have also unveiled the “Battle of 100 Models” for domestic large models. Because, other companies such as Baichuan Big Model, Zhipu Qingyan, and SenseChat were also opened for use on the same day.
Domestic large models enter the blowout era!
1. The battle for large models, a war without gunpowder
Large models, or large language models, refer to deep learning models trained using large amounts of text data, which can generate natural language text or understand the meaning of language text.
Simply put, large models can simulate the process of human language learning and understand and generate text in a human-like manner, which is an important way to artificial intelligence.
Siri and Tmall Genie, which we often use in our daily lives, are actually basic applications based on large models. However, because the AI of this period was "not intelligent enough", people nicknamed it "artificial retardation". However, no one has Thinking about it, in just a few years, AI has evolved from "understanding instructions" to "executing instructions", and can even learn by itself based on instructions to create copywriting, scripts, and drawings.
From the perspective of large model R&D, large models are mainly divided into three categories: one is self-developed by large manufacturers. For example, Baidu’s Wenxinyiyan, Alibaba’s Tongyi Qianwen, Huawei’s Pangu, Tencent’s Hunyuan, JD.com’s Yanxi, Byte’s Volcano Ark, Ant Group’s Zhenyi, Xiaomi’s MiLM-6B, etc. One type is an independent entrepreneurial team. For example, Sogou founder Wang Xiaochuan officially announced "Baichuan Intelligence" on the second floor of the Sohu Building in Wudaokou. There is also the academic group. For example, the MOSS large model released by the Fudan University School of Computer Science team at the beginning of this year, ChatGLM by Tsinghua University, and Zidong Taichu by the Institute of Automation of the Chinese Academy of Sciences.
In the capital market, large model manufacturers at home and abroad are also favored by capital.
After ChatGPT became popular, in just six months, Anthropic, an artificial intelligence company founded by the former OpenAI leader, received three rounds of large-scale financing, totaling more than US$850 million.
The number of large AI models released publicly in China has already reached hundreds. Counting companies with the AIGC label separately, financing transactions in the first half of 2023 were 5.89 billion yuan, and the number of incidents was 42, far exceeding previous years.
The big Chinese manufacturers with a keen sense of smell have even shown the courage to make a desperate move!
