In late 2016, Amazon Amazon Go was inaugurated in Seattle. This unmanned physical store, which makes full use of deep learning, sensors, computer vision and other AI technologies, had a huge impact on the retail world as soon as it appeared.
The industry changes brought about by AI don't stop there. Looking at the present, the growth dilemma faced by retailers has intensified their expectations for AI technology - objective factors such as the iteration of the economic cycle have made the consumer market riverbed bare, and business has returned to the essence of cost reduction and efficiency, and experience enhancement.
"The future will be a new, more efficient data processing system." Hraley, author of A Brief History of the Future, once predicted this about the possibilities of the digital future. A fully connected and digitized society may still be envisioned, but the relationship between people, the real world, and the virtual world is beginning to lap the shore in the waves rolled by generative AI.
After ChatGPT caused a wave of socialization, the industry has seen a "battle of a thousand models". However, beyond the changes perceived by the public in chatting, writing poems and painting ......, what kind of big model will be the key to the future industrial intelligence?
In the "2023 Jingdong Global Science and Technology Explorers Conference and Jingdong Cloud Summit", Jingdong speech rhinoceros big model, to give the answer to this question. This big model originating from the industry and serving the industry, taking the digital supply chain as the starting point, breaks the bottleneck of the application of the big model. The training fusion of 70% of the general data and 30% of the native data of the digital supply chain, is committed to knowledge-intensive, task-oriented industrial scenarios, to solve real industrial problems.
Social sound volume from the frontiers of science and innovation, generalized in the public carnival, with ChatGPT as the representative of generative AI, in the completion of the education market, shaping the cognition, the heat gradually faded. The objective value of solving the big model and scene severance lies in the AI anxiety implied in the heads of countless entrepreneurs.
According to SimilarWeb data, the growth rate of pre-ChatGPT's visits is amazing, the ring growth rate in January was 131.6%, in February was 62.5%, in March was 55.8%, and it slowed down significantly in April, with a ring growth rate of 12.6%, and in May, the number has changed to 2.8%, and it is expected that the ring growth rate in June is likely to be a negative number.
Big industry models that solve real problems in serious scenarios are starting to take the vocal high ground. Internationally, Microsoft, Amazon and other large manufacturers have begun to explore the path of commercialization of enterprise-class services; domestically, such as Baidu, Ali, Jingdong, Tencent, Huawei are accelerating the investment in the industry's large models, and the amount of mergers and acquisitions of enterprises of related technology routes has climbed to new heights.
Different vertical industries are calling for the emergence of intelligence to cope with the new technology, the call of the new era. The basis for the emergence of intelligence is to provide sufficient data and application scenarios for AI big models. For a single industry, this is not a one-day effort, but requires the ability of two aspects: a stronger data platform and smarter interaction capabilities, so as to obtain proprietary data as the "fuel" for the big model to the industry.
This has become the main theme of the company's recent technological exploration. The AI anxiety that has lingered for a long time is beginning to find an outlet. As far as industry players are concerned, a pioneering insight is the need to return the technology to the industrial scene itself, and actually enter the actual business flow, in order to solve the real industrial problems.
After the industry reached a consensus, it was found that a large model based on a generalized large model that incorporates proprietary data would be the last kilometer of the AI-enabled industry.
This verifies the judgment of Jingdong Cloud: intelligence, which will not be achieved overnight, will be prioritized in the digitization of leading industries to land. The retail industry, with its complex marketing scenarios and supply chain links, is an excellent "battlefield" for refining large models.
How to root into the long supply chain of the retail industry? How to let the data complete the whole chain of "immersion" from commodities to consumers? Many retail enterprises have returned to reality and reason, and they keenly realize that the high cost and technical barriers of general big models are not applicable to every enterprise, and that the industry big models that conform to the needs of actual scenarios are the wheels that have already been built.
Therefore, the big model of the competition, the key is not in technology, but in the industrial scene landing. This may also be, Jingdong launched industry-oriented speech rhinoceros big model, as well as the release of a series of products and solutions to promote the industrial landing in the industry after the cause of heated debate.
AI empowers the last kilometer of the industry, taking the lead in the retail industry to take a solid next step.