What AIGC technologies are already in use in manufacturing?

In addition to Sora, after the release of ChatGPT, have large models made any headway in manufacturing?

LDSTA-01  Lenovo Song Tao from the perspective of the distribution of intelligent computing resources told the number of intelligent front line, the current domestic demand for large models, 70% of the business in the Internet, 10% in scientific research units, fall in the manufacturing industry, the main opportunity is car autonomous driving.

Several senior figures in car companies told the number of intelligent front lines that in the past few years, the automotive industry has become more and more volled-up. The money of car companies is invested at both ends of the smile curve, one is the research and development end, and the other is the marketing service end. On the manufacturing side, over the past six months, there have been no typical scenarios. The main scenarios of the landing of the large model are focused on intelligent driving, marketing services and intelligent cockpit.

For example, data synthesis and data annotation in intelligent driving; In the after-sales service, when encountering some uncommon faults, the large model can find the situation that is closest to the description, quickly give the sort, assist the after-sales service personnel, and help the owner to investigate one by one, so that the after-sales service is highly acceptable.

LDSTA-01

One focus of competition this year is the smart cockpit. It is expected that in the first half of the year, some car companies will successively announce access to large models. However, the industry has not yet found an explosive scene in the smart cabin, and this year it will further explore the scene.

In 2024, the budget of car companies on large models is generally more than 10 million, which is used for autonomous driving, intelligent cockpit or large model private cloud deployment. That’s a significant increase from 2023.

In addition to the automotive industry, other intelligent manufacturing industries have made some progress in the research and production, supply and marketing of services.

Based on the “source” large model, they have trained in three directions LDSTA-01  of research and development. One direction is to train an assistant “Jarvis” for engineers. The other is an intelligent coaching system, a bit like the “iFlytek learning machine”, which takes engineers to learn and train repeatedly. There is also a direction of software design, through the automatic generation of code, to achieve the server related soft and hard links. At present, the internal testing of these directions has begun and is in the process of continuous optimization.