Professor Peking University HSBC keynote lecture
What is ChatGPT?
Why doesn't China have ChatGPT?
Can we "copy" ChatGPT?
How has our lives changed since ChatGPT?
With these questions
Let's take a look together
Professor Peking University HSBC keynote lecture
【Business Model Innovation and Application Opportunities in the Big Model Era】
Give us analysis and suggestions
Professor Wei Wei of Peking University HSBC Business School was specially invited to give a keynote lecture entitled "Business Model Innovation and Application Opportunities in the Era of Big Model (GPT)".
Wei Wei
Professor of Management, Peking University HSBC Business School
Director of the Business Model Research Center
Founder of "Wei Zhu Business Model Theory"
Inventory of the current situation of the era of large models
First of all, Mr. Wei popularized what ChatGPT is, as well as its characteristics and principles.
ChatGPT, full name is Chat Generative Pre-trained Converter, is the abbreviation of English Chat Generative Pre-trained Transformer, is an artificial intelligence chatbot program developed by the OpenAI Artificial Intelligence Research Lab. At present, it is still interactive in text mode, in addition to interacting in the way of human natural dialogue, it can also be used for more complex language work, including automatic text generation, automatic question answering, automatic summary and other tasks. Since the release of ChatGPT in November 2022, OpenAI valuation has risen to $29 billion. There were 1 million users in 5 days of launch, and hundreds of millions of users in two months of launch.
With the birth of ChatGPT, domestic and foreign companies have also laid out the field of large models, including Google released Bard in February this year, Microsoft released New Bing, and with the release of GPT-4 in March, Bloomberg, which provides global news, business and financial data services, launched BloombergGPT as a multinational mass media company. In the blink of an eye in China, Baidu took the lead in releasing Wen Xin's words at the end of March, followed by SenseTime's release of the new SenseNova, Alibaba's release of Tongyi Qianwen, and the upcoming Huawei Pangu, Tencent Mix, JD ChatJD, iFLYTEK cognitive intelligence model and so on.
Does the emergence of the same type of large models in various big manufacturers mean that we have successfully copied our own ChatGPT? Wei then put forward his own point of view: instead of asking why China does not have ChatGPT, we should ask why China does not have OpenAI.
After taking stock of the domestic large-language model industry value chain, the core technical personnel of large factories have switched to the second entrepreneurial tide of the big model business, and Mr. Wei analyzed the impact of AI on future careers through comparison and research of domestic and foreign data. At the same time, Mr. Wei also pointed out that the emergence of ChatGPT officially opened the fourth scientific and technological revolution in human history, and the eight major application directions of GPT have penetrated into our daily life scenes, "Each round of scientific and technological revolution will inevitably produce giant enterprises, and at the same time drive the development of the country and society." And with the opportunity comes people's anxiety, despair. ”
Advantages and disadvantages of ChatGPT
Through the study of the history of ChatGPT development, long-term practical experience and analysis of basic principles, Mr. Wei summarized the characteristics and shortcomings of ChatGPT, and clearly explained its advantages and disadvantages. After in-depth analysis, everyone clearly realized that the core advantage of large model technology is only to express different words in a logical and humanized way, but due to its centralized, integrated and closed-source business model characteristics, its limitations are also quite obvious. First of all, information is not always accurate, user data and privacy are not protected, data property rights are easily violated by centralized large model companies, and ultimately the creation of new data and new knowledge is greatly inhibited.
Know it, know why. What can we do to improve the shortcomings of existing large models? Are there more opportunities hidden here? Mr. Wei also revealed to us that PHBS ExpertGPT, developed by the team of China Enterprise Information Center of Peking University HSBC Business School, led by him, uses the high-accuracy professional knowledge map in the business field generated by small models to fine-tune and verify the open source large model twice, which can solve the inherent problems of fabricated knowledge, data property rights, privacy leakage and incentive problems generated by new knowledge in centralized large models. "China should seize the development opportunity of a new round of scientific and technological revolution, develop a multi-center distributed technology architecture of generative AI + discerning AI, and accelerate the global catch-up of AI technology."
Big model business model innovation and competition
Finally, back to the topic of business model innovation and design that Mr. Wei is best at. He introduced the business model of the integrated large model symbiote, as well as the innovation and thinking of various business models. In view of the future competition pattern of the big model, Professor Wei believes that there is both cooperation and competition between China and the United States under the situation of separation. In view of the different heritage and cultural traditions, the business models preferred by large enterprises in the two countries are also different, with US companies preferring the same model to meet the same needs of different customers, while Chinese companies are better at building business models around the different needs of the same customers, which directly leads to the different scale distribution of large model manufacturers in the future. But no matter what the development is, we should learn and understand the innovation and opportunities of large-model business models with a more positive attitude, rather than directly mything the impact and progress brought by AI. Because before the advent of Chat-GPT, most people didn't know what OpenAI did and how much it spent.
The audience said that the lecture brought them a lot of cutting-edge information, and they were deeply inspired, and they also lamented that the future development of domestic AI models is heavy and long.