How did DeepSeek shock the world?

| Read time: 5 minute(s)
How did DeepSeek shock the world? 27th January 2025 may go down in history. Why? The sudden appearance of DeepSeek – a chinese owned AI chatbot app that rivals western inventions like ChatGPT. It sent the stock markets revealing. Politicians reacted with concern about an AI-model that spews the Chinese Communists Party’s official line. Deepseek makes an app similar to chatGPT OR Claude. It will answer questions, or script an empathetic break up text but it distinguishes itself by articulating its reasoning before responding to a prompt and what’s extraordinary is its apparent quality and low cost and ofcourse its seeming appearance out of nowhere. DeepSeek is a Chinese AI startup that was formed in 2023 and spun out of a hedge fund. It was the product of 40 year old entrepreneur Ieang W Fang. Well western tech leaders are extremely good at self promotions, he’s kind of stayed out of the spotlight. He studied electrical engineering and computer vision which is a very interesting field of AI specifically. To the AI world DeepSeek wasn’t completely new but in November it released its preview of the R1 model and later that year we could actually see that it was performing really well against some benchmarks. It quickly surged to the top of the apple app store and surpassed ChatGPT. Just don’t ask questions about Chinese politics. It’s reasoning is aligned with that of Beijing’s sensors meaning some questions are politely sidestepped. US app perplexity has already announced that DeepSeek can be used uncensored on its platform but what really sent investors over the edge is that DeepSeek revealed that it was able to build model and train it at just a fraction of the cost. The fact that DeepSeek was able to do this really raises the question - what have these companies been spending billions of dollars on? DeepSeek offers several advantages over other AI models, including strong reasoning and logic capabilities, robust multi-language support, a focus on explainability, and open-source accessibility. DeepSeek models, particularly R1, are known for their performance in complex reasoning tasks and logical deduction. They also excel in tasks requiring mathematical reasoning, surpassing models like OpenAI's o1. Furthermore, DeepSeek's open-source nature allows for customization and adaptation to specific needs. This has been a wake up call for AI industry since the US has spent a long time underestimating China. China’s tech sector is very capable of innovating and creating these very powerful and very advanced products and about time Silicon Valley should not take China for granted.

27th January 2025 may go down in history. Why? The sudden appearance of DeepSeek – a chinese owned AI chatbot app that rivals western inventions like ChatGPT. It sent the stock markets revealing. Politicians reacted with concern about an AI-model that spews the Chinese Communists Party’s official line.

Deepseek makes an app similar to chatGPT OR Claude. It will answer questions, or script an empathetic break up text but it distinguishes itself by articulating its reasoning before responding to a prompt and what’s extraordinary is its apparent quality and low cost and ofcourse its seeming appearance out of nowhere. DeepSeek is a Chinese AI startup that was formed in 2023 and spun out of a hedge fund. It was the product of 40 year old entrepreneur Ieang W Fang. Well western tech leaders are extremely good at self promotions, he’s kind of stayed out of the spotlight. He studied electrical engineering and computer vision which is a very interesting field of AI specifically. 

To the AI world DeepSeek wasn’t completely new but in November it released its preview of the R1 model and later that year we could actually see that it was performing really well against some benchmarks. It quickly surged to the top of the apple app store and surpassed ChatGPT. Just don’t ask  questions about Chinese politics. It’s reasoning is aligned with that of Beijing’s sensors meaning some questions are politely sidestepped. US app perplexity has already announced that DeepSeek can be used uncensored on its platform but what really sent investors over the edge is that DeepSeek revealed that it was able to build model and train it at just a fraction of the cost. The fact that DeepSeek was able to do this really raises the question – what have these companies been spending billions of dollars on?

DeepSeek offers several advantages over other AI models, including strong reasoning and logic capabilities, robust multi-language support, a focus on explainability, and open-source accessibility. DeepSeek models, particularly R1, are known for their performance in complex reasoning tasks and logical deduction. They also excel in tasks requiring mathematical reasoning, surpassing models like OpenAI’s o1. Furthermore, DeepSeek’s open-source nature allows for customization and adaptation to specific needs.

This has been a wake up call for AI industry since the US has spent a long time underestimating China. China’s tech sector is very capable of innovating and creating these very powerful and very advanced products and about time Silicon Valley should not take China for granted.

China

Rate Now

Read to me