DeepSeek: China’s New AI

Lesson made on February 18, 2025

Practice the Vocabulary in this Lesson


Download Printable Lesson (PDF)

Download Printable Vocabulary Practice (PDF)

Download Printable Answer Key (PDF)

Reading Lesson:

DeepSeek is a Chinese AI company that became famous in early 2025 after launching an advanced AI model called DeepSeek-R1. Just weeks after its release, the app became the most downloaded on Apple’s U.S. store, and its success shook the American tech industry. Many U.S. technology companies, especially NVIDIA, saw their stock prices fall dramatically as DeepSeek’s AI was far cheaper and more efficient than its American competitors.

DeepSeek’s AI models stand out because they use much less computing power while still producing high-quality results. Unlike American companies that rely on huge amounts of data and expensive hardware, DeepSeek has found new ways to make AI work efficiently. It also took an unusual step by making its technology open to the public, while U.S. companies usually keep their AI systems secret.

The company was started by Liang Wenfeng, a math expert who started using AI for stock trading before shifting his focus to artificial intelligence research. By 2021, he had bought enough powerful NVIDIA chips before the U.S. government limited their sale to China. This early planning helped DeepSeek grow despite restrictions on advanced computer hardware.

DeepSeek’s success has intensified competition between China and the U.S. in artificial intelligence. The U.S. has tried to block China’s access to advanced AI chips, but DeepSeek has proven that China can make AI technology with fewer resources by using more efficient methods.

Comprehension Questions:

1. Why are American technology companies worried about DeepSeek?

a) DeepSeek’s AI is less powerful.
b) DeepSeek’s AI is cheaper and more efficient.
c) DeepSeek is not available in the U.S.
d) American companies control DeepSeek.

2. NVIDIA saw its stock prices fall dramatically. What does “fall dramatically” mean?

a) Decrease suddenly by a large amount.
b) Decrease slowly over a long period of time.
c) Stay the same.
d) Change in a small way.

3. Why did DeepSeek’s popularity affect NVIDIA’s stock price?

a) NVIDIA owns DeepSeek.
b) DeepSeek’s AI did not need as many expensive chips.
c) DeepSeek bought all of NVIDIA’s chips.
d) NVIDIA stopped making computer chips.

4. Which of these statements is false?

a) DeepSeek-R1 became the most downloaded app in the U.S.
b) DeepSeek’s success caused American tech stocks to fall.
c) The U.S. government restricted sales of NVIDIA chips to China.
d) DeepSeek’s technology requires more computing power than American AI models.

5. What does DeepSeek’s success suggest about AI development?

a) Expensive technology is always necessary.
b) Only American companies can create strong AI models.
c) AI can be improved with smarter methods, not just better hardware.
d) The U.S. controls all AI technology.

6. What was unusual about DeepSeek’s approach to AI development?

a) They spent more money than American companies.
b) They used more data than anyone else.
c) They made their technology open to the public.
d) They focused only on stock trading.

7. How was DeepSeek able to keep developing AI despite U.S. restrictions?

a) The U.S. gave them special permission.
b) They stopped using AI chips.
c) They moved their company to the U.S.
d) They purchased advanced chips before restrictions began.

8. Which is the correct order of events?

a) DeepSeek launched AI → U.S. limited chip sales → NVIDIA’s stock fell
b) U.S. limited chip sales → DeepSeek launched AI → NVIDIA’s stock fell
c) NVIDIA’s stock fell → U.S. limited chip sales → DeepSeek launched AI
d) DeepSeek launched AI → NVIDIA’s stock fell → U.S. limited chip sales

9. What might be a long-term effect of DeepSeek’s success?

a) The U.S. might lose its leadership in AI technology.
b) AI will stop developing.
c) All companies will use American technology.
d) AI will become more expensive.

10. What is the main idea of the article?

a) DeepSeek’s AI is the most expensive in the world.
b) The U.S. government limited the sale of NVIDIA chips to China.
c) There are many technology companies that are working on artificial intelligence.
d) DeepSeek has developed an efficient AI that challenges American companies.

Speaking and Writing Activities:

– Discuss these questions with a partner or a small group.

– Choose one topic and write a response to it. Show your writing to a classmate or teacher.

1. If you had to explain AI to a friend, how would you describe it in simple words?
   
   
2. What effects will AI have on the world? Will these effects be good or bad?
   
   
3. Would you rather work for a big company or start your own small business? Why?

More Reading Practice:

NVIDIA

NVIDIA is a technology company known for creating products that help computers perform better. They design and produce graphics processing units, or GPUs, which are important parts of a computer that help it display images on screens. These GPUs are used in many devices, such as gaming computers, workstations, and even self-driving cars. In addition to making hardware, NVIDIA also develops software that allows computers to run more efficiently.

NVIDIA is especially well-known in the world of gaming. Gamers use its graphics cards to make sure their games look smooth and realistic. The company also plays a key role in artificial intelligence (AI), where its technology is used in fields such as healthcare, robotics, and finance. They are recognized for their innovations in AI, which can help machines learn and make decisions on their own.

The company was founded in 1993 by Jensen Huang, Chris Malachowsky, and Curtis Priem. Its headquarters are in Santa Clara, California. Over the years, NVIDIA has grown into one of the leaders in the tech industry, continually developing new products that improve computing technology.

DeepSeek-R1

DeepSeek-R1 is an advanced artificial intelligence (AI) model developed by the Chinese company DeepSeek, which was started in July 2023. It is designed to perform complex tasks such as mathematical problem-solving, coding, and logical reasoning. This model has gained attention for its high performance and cost-effectiveness.

One of the key features of DeepSeek-R1 is its open-source nature, which means that its code is freely available for anyone to use and modify. This openness promotes collaboration and innovation within the AI community. Additionally, DeepSeek-R1 is known for its efficiency, achieving results comparable to other leading AI models at a fraction of the cost. For example, it has been reported that DeepSeek-R1 was trained at a cost of approximately $6 million, significantly less than the $100 million spent on training OpenAI’s GPT-4 model.

In terms of performance, DeepSeek-R1 has demonstrated impressive capabilities. It achieved a 97.3% accuracy rate on the MATH-500 benchmark, indicating strong proficiency in mathematical reasoning.

DeepSeek-R1 has been integrated into various platforms and services, including services offered by Microsoft, Google, and Amazon.

ESL Lesson about Elon Musk

ESL Lesson about Neuralink

ESL Lesson about Chinese animated movie “Ne Zha 2”

ESL Lesson about a Chinese influencer who died from alcohol poisoning in a live video