What should the manufacturing industry do to prepare for the large model?

Starting in 2024, almost every week, big models have had “breaking” news launched. According to the industry’s information, in the next, there are still some companies released. Many companies contacted by Peng Xinyu, vice president of Alibaba Group and CEO of Hillhouse Sheep, are anxious that they do not know what to do. Under the continuous extension of anxiety, some social anomalies such as “the end of AI is to sell lessons” have also appeared in society.

MSK070C-0150-NN-S1-UG0-NNNN Lu Chong admitted that the concept is now much talked about, but the real application needs to be done step by step.

“If it comes down to the enterprise level, there are three things: computing power, the data of the enterprise, and the talent of the enterprise.” Peng Xinyu analysis said that in terms of computing power, in addition to a few companies such as OpenAI, for most companies, everyone’s starting line is almost the same.

In terms of data, OpenAI has basically collected all the public data on Earth, but the most valuable data for enterprises is often the enterprise’s own data, which is not collected by OpenAI. Enterprises need to consider how to improve the quality of this data, turn it into a good MSK070C-0150-NN-S1-UG0-NNNN asset, and integrate it with the larger model. “I think it’s a real investment that companies can hold onto.”

In terms of talent, for most enterprises, it is not necessary to find a bull in the field of AI, and the key is that such talents can not be found now. Instead, companies need people who understand how to apply big models, how to talk to enterprise workflows, and translate the power of big models into business productivity. “Companies need to catch these people.”

Mr. Wang emphasized the power of industry experts. He took live delivery as an example. The rise of social platforms has liberated individual productivity and changed the operation mode of many traditional industries. But the success of this model often depends on experts who have a deep understanding of a particular industry.” For example, an expert who is well-versed in selling cosmetics may not be good at live-streaming sales of electronics, and vice versa. He said that while technology offers varying degrees of new opportunities in various fields, true expertise and industry understanding are still indispensable.

“No matter what form the big model takes in the future, it will be there.” ‘Companies need to be prepared,’ Mr. Peng said.

“Artificial intelligence is what level 5 of intelligent manufacturing can achieve.” Lenovo Song Tao said that enterprises need to begin to implement digital work such as basic business combing, software applications and platforms, and lay a good foundation. These also need the support of infrastructure, cloud computing, hybrid cloud, data security, data management and other basic work, which is also an opportunity to help enterprises re-organize IT, OT, DT, ET and other opportunities.

Song Tao also mentioned the four elements of artificial intelligence MSK070C-0150-NN-S1-UG0-NNNN  technology landing – computing power, data, algorithms and scenes. Among them, the scene is the driver, and it is necessary to excavate scenes that can use AI to reduce cost and improve efficiency. Computing power is the underlying support, data is the nutrient, and algorithms are the value formed by artificial intelligence technology.

He suggested that in the implementation of artificial intelligence, priority should be given to the construction of computing power and data storage, and then data governance and data value mining; Use high-quality data to train the corresponding algorithms, so as to generate value for the business. The landing of AI also needs corresponding talents, and it is realized step by step and by stage through technology.

Peng Xinyu referred to the “development of new quality productivity” in the government work report last week. “Large models for smart manufacturing are desirable.” He said that from what OpenAI has done this year, it has opened up new Spaces – human space, economic space, and enterprise space. “In essence, we haven’t seen real disruption in the industry yet, but it’s also the responsibility of these people closer to the corporate scene.”