Meenu
A Chinese AI DeepSeek has shaken the world. American tech companies and the US stock market have been shocked to the extent that they couldn’t have ever imagined. On 20th January 2025, a Chinese research lab launched its AI chatbot. It’s named DeepSeek R1, and they publish the research paper along with it. They stated that this chatbot is much better than most advanced chatbots across benchmarks like maths and reasoning. This includes OpenAI’s ChatGPT o1 model, Meta’s Llama, and Google’s Gemini Advanced; it has overthrown all of them.
DeepSeek is the best in terms of performance. It’s the most efficient and takes only a fraction of the time and money to develop it. The most fantastic thing is that using it’s completely free of cost for all of us. On the other hand, OpenAI charges $200 per month to use its ChatGpt o1 Pro model. Also, the cost of training DeepSeek is reported to have been only $5.6 million. Whereas other American companies, whether it’s OpenAI, Meta, or Google, are spending billions of dollars to develop their AI models.
Coincidentally, a while back, OpenAI’s founder Sam Altman was asked whether foundational AI models like ChatGPT can be made by Indians in India. Then, he answered arrogantly that no one could do it except them. People can try, but he said it would be ‘hopeless.’ Today, Sam may be feeling hopeless because just a week after DeekSeek’s launch, it became the most downloaded app in the US on the App Store and Google Play Store. It has left behind ChatGPT. It became the number one app in the App Store in India and other countries. And by 27th January 2025, it has disrupted the American financial markets.
The Downfall of NVIDIA
NVIDIA was the most valuable company in the world before the launch of DeepSeek AI, valued at $3.5 trillion. But in just one day, it dropped to $2.9 trillion. This is a computer chip manufacturing company that specializes in making computer chips that can be used to train and operate AI. NVIDIA’s shares fell 17% in a single day, and its valuation lost $589 billion. This is the biggest loss faced by any company in history. The benchmark of US tech companies, Nasdaq, dropped 3.1%
. But why did NVIDIA suffer such a significant loss specifically? We’ll talk about the interesting reason later in this article.
Before that, let’s find out the story behind DeepSeek. Where did this AI come from, and why has it shocked the world?
What is The Story Behind DeepSeek AI?
The credit for developing DeepSeek goes to a 40-year-old Chinese entrepreneur, Liang Wenfeng. He rarely makes public appearances and prefers to keep his identity hidden. There’s not much information about his life and history available to the public. In 2015, he founded a hedge fund called High Flyer. It uses mathematics and AI to make investments. In 2019, he founded High Flyer AI to research AI Algorithms. In May 2023, he used the money he was earning from his hedge fund to start a side project of creating an AI model.
Liang said he wanted to create his own AI model, which would be better than all the existing AI models worldwide. He aimed to satisfy his scientific curiosity. He didn’t seek to make money and earn lots of profits. When he started to gather his research team to make his AI model, he didn’t hire engineers. He hired PhD students from China’s top universities. He used the most difficult questions around the world to train his AI model. And within just 2 years, and after spending a few million dollars, he launched his DeepSeek R1 model. Only 200 people were involved in making it, and 95% were under 30 years old.
Architecture of DeepSeek AI
It’s quite interesting to understand the architecture of DeepSeek. It’s a Chain of Thought Model. OpenAI’s ChatGPT o1 model is the Chain of Thought Model, too. The older AI models had considerable shortcomings. However, the Chain of Thought process has now removed this drawback. That’s why models like o1 and DeepSeek are called Advanced Reasoning Models. They’re better at logically answering questions. Another positive aspect of DeepSeek is that it shows you exactly what it’s thinking, step-by-step and in great detail, before answering.
Downside of DeepSeek AI: Censorship
After DeepSeek’s release, its biggest downside is its censorship. Let’s understand this with an example:
Ask any critical questions related to the Chinese government and politics. Such as
- “What happened at Tiananmen Square in 1989 in China?”
- “What is the biggest criticism against Xi Jinping?”
- “Why is China criticized for human rights abuses?”
DeepSeek will give you the same answer to all these questions, i.e., “Sorry.” “Sorry, that’s beyond my current scope. Let’s talk about something else.”
But if you ask DeepSeek to criticize any other world leader, whether it’s Joe Biden, Donald Trump, or Putin, it answers in great detail.
The Truth!
Actually, the AI models developed in China ask around 70,000 questions as a test to check whether they will answer politically controversial questions safely. Only after this are these Chinese AI models unable to answer such questions. China’s Chief Internet Regulator, the Cyberspace Administration of China, tests them.
DeepSeek R1 AI: An Open-Source Software
Now, some people say that we should boycott it completely because its answers are full of Chinese propaganda. But friends, the important thing here is that DeepSeek is an open-source software. This means its source code is publicly available. Anyone can download it locally.
So, one way to use DeepSeek is to download its app from the App Store or Google Play Store. Another way is to download its entire code and run this AI locally on your system. By doing this, you can change its code and modify it yourself according to your use case.
Other American companies have already begun doing this, such as Perplexity AI. They downloaded the DeepSeek R1 model and erased all censorship, and now you can use the R1 model in Perplexity Ai. Microsoft did this, too. The DeepSeek R1 model is now available on Azure AI Foundry. They have also announced that users will soon be able to use this model in Copilot.
People have used DeepSeek’s open-source function to mock American companies. Organizations like OpenAI, which has the word ‘Open’ in its name, keep its source code close. Initially, these companies promised to work for the public and keep everything open source. But in reality, they weren’t the same.
Elon Musk has often trolled OpenAI, calling it ClosedAI. Today, this Chinese AI is more open than OpenAI.
Artificial Analysis of All AI Chatbot Models
Let me show you a performance comparison among the top AI chatbots.
OpenAI’s ChatGPT, Anthropic’s Claude, Google’s Gemini, Alibaba, another Chinese company’s Qwen 2.5, Meta, Facebook’s parent company’s Llama. If all these are compared with DeepSeek, what would be the results?
According to Artificial Analysis, DeepSeek AI is at the top in terms of coding. Followed by ChatGPT, Claude, Qwen 2.5, and finally Llama.
DeepSeek is at the top in quantitative reasoning, followed by ChatGPT and Qwen 2.5.
In terms of scientific reasoning and knowledge, ChatGPT is at the top, followed by DeepSeek AI, Claude, and finally, Llama.
The Artificial Analysis website has rated ChatGPT o1 on Quality and DeepSeek R189.
When PCMag tested AI chatbots, they found the following:
- DeepSeek AI is the best in terms of news-related knowledge.
- In calculations, ChatGPT and DeepSeek are equal.
- In terms of writing poems, ChatGPT wins.
- As for creating tables, ChatGpt is good.
- In solving riddles, DeepSeek AI is the best.
Disadvantage for DeepSeek AI
A significant disadvantage for DeepSeek is its response time. Look at the following graph showing latency. The number of seconds it takes to get a response after asking a question. o1 takes 0.7 seconds, and DeepSeek takes 3.6 seconds.
In recent days, this response time has actually increased because DeepSeek has become so popular all over the world that everyone wants to download and use it. Because of this, they’re facing issues with their servers. They run busy, so it starts taking much longer to respond.
So, this is a significant disadvantage.
OpenAI ChatGPT o1 vs. DeepSeek R1 AI
If you compare ChatGPT o1 and DeepSeek to a large extent, DeepSeek is much more innovative than o1. One of the examples is its Mixture of Experts Method.
Mixture of Experts Method is in Chatbots
When you ask ChatGPT o1 for something, it works as a single model. Whatever your question is, ChatGPT will be your engineer, doctor, and lawyer. However, DeepSeek has divided itself into multiple specialized models. DeepSeek AI has separate engineers, doctors, and lawyers. Depending on your question, only the engineer or the doctor would be called.
With this, the time taken for data transfer is reduced. Second, the number of parameters that must be turned on is also reduced.
Did DeepSeek’s Source Code Copied from OpenAI?
Another criticism against DeepSeek is that they have copied their AI model from OpenAI. OpenAI claims that there is evidence showing that DeepSeek has used its proprietary models to train itself. They claim to have distillation evidence, where the output from bigger models is used to improve the performance of smaller AI models. This helps smaller models give the same results at lower costs.
The Game Backfired on OpenAI Itself!
To some extent, ChatGPT has copied things from all over the internet to train itself. Many books were used without the permission of the authors to train their models.
That’s why 17 renowned writers, including the Game of Thrones author George RR Martin, filed a copyright infringement suit against OpenAI in September 2023. The New York Times has also filed a case against OpenAI and Microsoft.
Other than this, 8 American newspapers and many Canadian news outlets have filed cases on OpenAI.
The Inception of AI War!
As users, this is good for us, but we’re seeing the beginning of AI wars at the country and company levels. In 2022, the American government imposed export control so that other countries cannot use the computer chips needed to train AI—especially Chinese AI companies. It included NVIDIA’s H100 chips; Chinese companies were unable to buy them. This was problematic because they had to use NVIDIA’s older computer chips to train their AI models.
America tried to prevent other countries from developing foundational AI models. They did this by denying them access to computer chips needed for training AI models. So, the employees working at DeepSeek were forced to innovate. They created software that uses only a fraction of resources and works more efficiently at only a fraction of the cost.
Friends, that’s why NVIDIA’s stock fell the most. A few months ago, companies like Meta, OpenAI, and Google claimed that we would need more chips, energy, and money if we wanted to scale AI to a higher level. But DeepSeek has done this using less money and energy and outdated computer chips.
Summing Up!
Finally, you can see that anyone can take advantage of this opportunity to upskill themselves in the field of AI. Undoubtedly, those who ignore AI now will be left behind in the future. It isn’t easy to imagine how impactful AI will be in the future, globally, in every sector, in every field. So, let’s keep updating and see how much more AI will offer to the world!
GIPHY App Key not set. Please check settings