"Consensus drives shared benefits and good governance promotes intellectual advancement": The 2024 World Artificial Intelligence Conference and Global Governance High-Level Forum on Artificial Intelligence (WAIC) was held in Shanghai from July 4th to July 6th. Hanwang Technology Co., Ltd. (referred to as "Hanwang Technology") is a high-tech enterprise focusing on AGI general artificial intelligence and robotics. Faced with the impact of AI technology on the consumer electronics industry, Wang Jie, vice president of Hanwang Technology, did an interview with the reporter of "Star Daily" in the Caixin Media live broadcast room during the conference.

Recalling the development background of Hanwang's "Tian Di Model", Wang Jie emphasized that Hanwang Technology has always been committed to human-machine interaction intelligent technology. Handwritten recognition and OCR recognition are part of human-machine interaction. All human-machine interactions require input and output, just like electronic paper products, which completes the complete closed loop of displaying some knowledge to users through electronic paper products - simulating brain understanding - output. This is called human-machine interaction.

"Hanwang has always been doing NLP, but with the current blessing of the large model and the combination with the new trust model, it should fill the last" short board "". This is an opportunity to provide users with more comprehensive intelligent product services for human-machine interaction."

When discussing how physical products such as Hanwang's handwritten e-books can gain market competitiveness in the future, Wang Jie believes that the future environment is intelligent, and different interaction modes will have different interactions under the overall framework. They coexist. Among them, the devices we are now focusing on, such as electronic paper books, are biased towards the display field. They are eye-friendly, paperless, and environmentally friendly, and still have a place in the future.

Under the combination of the large model and the electronic paper book, office efficiency can be effectively improved and mutual conversion between multimodal content can be achieved. For students, the personalized dialogue of the large model and targeted answers can truly realize the concept of "teaching according to students' aptitudes" in the education field. Electronic paper books are no longer just "notebooks" in the traditional sense.

"In the future, paperless and intelligentization should be an inevitable road," said Wang Jie.

Regarding the issue of the application of the large model landing that is currently being heatedly discussed, Wang Jie said that the company is constantly laying out, and the digital management of libraries and archives can be made intelligent through the large model upgrade. And the large model has landing in office, education, courts and other scenarios. "We are focused on user and service orientation, using technology to provide customer services that are more targeted, deeper, and more accurate. This is our current direction."

The challenge of the large model to traditional manufacturers still requires more computing power to support, and cloud services are needed to provide services. It is still difficult to make a small model on the end side, but in the future development, Hanwang Technology focuses first on specialization - focusing on providing vertical content in a certain field, and second is small size. The size of the model is not the key point. The key is whether it can be applied in practice. The large model of artificial intelligence will eventually return to various industries.