CCTV News: Recently, the usage of DeepSeek has increased rapidly, and has become the fastest application to exceed 30 million daily active users. The reporter learned from the Ministry of Industry and Information Technology that the three basic telecommunications companies have fully connected to the domestic open source big model DeepSeek. What can I do after accessing?

Tianyi Cloud, a subsidiary of China Telecom, is the first operator-level cloud platform center in China to implement the implementation of DeepSeek's domestic inference.

After this video conferencing system is connected to DeepSeek, the meeting content, including the voice speeches of participants and the PPT content of the demonstration, can be synchronized to form meeting minutes and summarized into a well-organized summary.

China Mobile's cloud platform has specially customized exclusive computing power solutions for the popular DeepSeek-R1 model, providing cost-effective computing power support and development environment for Internet companies' innovation and university research.

On China Unicom's cloud platform, the reporter saw that the programming assistant could quickly generate more than a dozen lines of code with a single Chinese instruction, and the DeepSeek model can also provide detailed explanations and optimizations of the generated code.

Experts pointed out that when telecom operators access DeepSeek, they will make full use of their powerful network infrastructure and massive user data, deeply explore the potential of the DeepSeek big model, provide users with richer intelligent application scenarios, and promote the intelligent upgrade of the entire communications industry. Terminal manufacturers are accelerating their adaptation to DeepSeek

In addition to the operator's cloud platforms, terminal manufacturers such as mobile phones and PCs are also actively embracing DeepSeek, and some local governments have also begun to deploy DeepSeek in government affairs systems.

This is the first Android phone equipped with the DeepSeek model. According to technicians, in addition to the significant improvement in AI task processing efficiency, access to DeepSeek will also bring users more intelligent interactive experiences such as in-depth thinking and personalized services.

Industry insiders said that DeepSeek's open source model has greatly reduced the threshold for artificial intelligence to enter various industries, and related government and enterprise-level applications are expected to experience fission growth.

It is understood that the Shenzhen Longgang District Government Services and Data Management Bureau has deployed and launched the Deepseek-R1 full-size model on the government affairs external network, becoming the first government department unit in Guangdong to deploy the model in the government affairs information innovation environment. At present, the capabilities of the DeepSeek model have been opened to all departments in Longgang District. DeepSeek promotes the transformation and upgrading of chips and intelligent computing centers

The reason why DeepSeek is sought after in the industry is mainly due to its innovative algorithms and high cost performance. It not only optimizes the use of computing resources, but also brings new opportunities to the development of domestic chips and promotes the industry to move towards efficient and green directions.

Experts pointed out that the biggest advantage of DeepSeek is its high cost-effectiveness. As a super intelligent "brain", it is not only smart, but also very cost-effective to use.

DeepSeek has greatly reduced the cost of large-scale model training. Large-scale model training no longer requires so many high-end chips and data centers. Will related industries face certain challenges?

Experts said that in the short term, some cloud service providers may experience excessive computing power due to the lightweight model. But in the long run, the popularity of AI applications will promote the upgrade of smart computing centers to the direction of "efficiency + green". How to use the effectiveness of big models? Professionals share tips

In the era of big models, AI technology has moved from the laboratory to daily life and has become a right-hand assistant to work and life. How can we apply big models to fully utilize the effectiveness of AI?

First of all, you must ask your own questions accurately, so that you can achieve efficient interaction. Communicating with big models and asking questions is a technical job. For example, if you want the big model to help you plan your trip, don’t just say “write me a travel guide”, and provide specific details. For example, "I plan to go to Changsha for 5 days next month, with a budget of 5,000 yuan. I like food and natural scenery, so help me plan my itinerary."

Secondly, for large models, you need to maintain continuous optimization. Taking translation as an example, articles translated for the first time in the big model may have insufficient sentence fluency. Users can further optimize the answers through feedback. For example, you can point out that "the sentences here are not very natural, can you change to a more fluent expression?"

And another thing is very important, you must use prompt words skillfully. Prompt words are "spells" that communicate with the big model. Different prompt words can guide the big model to produce completely different output results. If you want to write a science fiction novel, you can use prompt words such as "Star Wars", "Extra Space", and "Aliens", and you can output more exciting works with detailed prompts.