DeepSeek: Everything you need to know about the AI ​​chatbot app

Deepseek has gone viral.

China’s AI Lab Deepseek has invaded mainstream consciousness this week That chatbot app has risen to the top of the Apple App Store chart (The same goes for Google Play.). Deepseek’s AI models were trained using computationally efficient techniques. He leads Wall Street analysts – and engineers – Question whether the US can maintain its lead in AI races and whether demand for AI chips will be maintained.

But where did Deepseek come from, and how did it quickly rise to international fame?

The origins of Deepseek traders

Deepseek is supported by Highflyer Capital Management, a Chinese quantitative hedge fund that uses AI to inform trading decisions.

AI enthusiasts Liang Wenfeng He co-founded High Flyer in 2015. Zhijiang University students reportedly launched High-Flyer Capital Management as a hedge fund in 2019, focusing on the development and deployment of AI algorithms.

In 2023, Highflyer launched DeepSeek as a lab dedicated to researching AI tools separate from the financial business. With a high flyer as one of the investors, the lab ran to its own company, also known as Deepseek.

From day one, Deepseek has built its own data center cluster for model training. But like other AI companies in China, DeepSeek is affected by the US export ban on hardware. To train one of the more recent models, the company was forced to use the H100’s Nvidia H800 chip, a less powerful version of the chips available to US companies.

The Deepseek technical team is said to distort young teams. company Reportedly, they are actively recruiting PhD researcher at a top university in China. Deepseek also hires people without computer science background According to the New York Times, the technique will help you better understand a wide range of subjects.

A powerful model from Deepseek

Deepseek announced a set of models for its first model (Deepseek Coder, Deepseek LLM, Deepseek Chat) in November 2023.

The DeepSeek-V2, a general purpose text and image analysis system, worked well on a variety of AI benchmarks and was much cheaper than comparable models of the time. It forced Deepseek’s domestic competition, including bytedance and Alibaba, to lower the usage prices of some models and make others completely free.

deepseek-v3It was released in December 2024 and added to the notoriety of Deepseek.

According to DeepSeek’s internal benchmark tests, the DeepSeek V3 outweighs both downloadable and openly available models like Meta Llamas Like Openai, “closed” models that can only be accessed via APIs GPT-4O.

Equally impressive is Deepseek’s R1 “Inference” model. Deepseek, released in January, claims R1 works similarly to Openai’s O1 model on key benchmarks.

R1 is an inference model, so there is effectively a check itself. This helps avoid some of the pitfalls that usually make the model trip. Inference models take a little time to reach the solution compared to regular irrational models. The advantage is that it tends to be more reliable in domains such as physics, science, and mathematics.

However, the other models of the R1, Deepseek V3 and Deepseek have their drawbacks. They are the AI ​​developed by China, so they are the target of it. benchmark China’s internet regulators ensure that their responses “embodies core socialist values.” For example, in Deepseek’s Chatbot app, R1 does not answer questions about Tiananmen Square or Taiwan’s autonomy.

A destructive approach

If DeepSeek has a business model, it is not clear what exactly that model is. The company offers products and services far below market value and offers them free of charge to others. Also, they have not received the money from investors.Despite many VC interests.

In the way Deepseek said, efficiency breakthroughs have allowed us to remain competitive at extreme cost. Some experts Conflict However, the numbers provided by the company.

In any case, the developers are using the DeepSeek model. This is not open source because the phrase is generally understood, but is available under a generous license that allows for commercial use. According to Clem Delangue, CEO of Hugging Face, one of the platforms that host Deepseek’s models The embracing face developer has created over 500 “derivatives” models of the R1 It was won by combining 2.5 million downloads.

Deepseek’s success against a larger, more established rival It is called “maintaining AI.” and “It’s exaggerated.” The company’s success was at least partially responsible Nvidia’s stock price falls 18% January, and Evokes public responses From Openai CEO Sam Altman.

Microsoft DeepSeek has announced that it will be available for Azure AI Foundry ServiceMicrosoft’s platform that connects AI services for businesses under a single banner. When asked about Deepseek’s impact on Meta AI spending during first quarter revenue calls, CEO Mark Zuckerberg said Expenditure on AI infrastructure continues to be a “strategic advantage” For meta. March, Openai, known as deepseek, is called “national aid” and “national management.” We recommend that the US government consider banning models from Deepseek.

During Nvidia’s fourth quarter revenue call, CEO Jensen Huang highlighted Deepseek’s “great innovation.” That and other “inference” models say it’s great for Nvidia because they require more computation.

at the same time, Some companies ban DeepseekAnd the whole thing is country and government, Including Korea. New York State too Deepseek has been banned from being used on government devices.

It’s not clear how Deepseek’s future will keep it. An improved model is given. But it appears that is the case with the US government Beware of what it perceives as harmful foreign influences. In March, the Wall Street Journal reported it The US may ban government devices from Deepseek.

This story was originally published on January 28th, 2025 and will be updated regularly.