A Chinese AI company, competing with ChatGPT, is becoming popular in Silicon Valley as it quickly grows and almost beats top American AI firms like OpenAI and Meta.
DeepSeek is a Chinese AI business that makes open-source large language models (LLMs), according to its website.
The company introduced R1, a specific model for handling difficult problems, on January 20. It quickly ranked in the global top 10 for performance and was built much faster and cheaper than other U.S. models, using fewer and less powerful AI chips, according to the Wall Street Journal.
The latest version of the app was announced on Inauguration Day for President Donald Trump, while TikTok, another social media app from China, was in the news over the possibility of being banned in the U.S.
Meta’s Chief AI scientist, Yann LeCun, shared his thoughts on social media about the app and how quickly it has become successful.
In a post on Threads, he noted that what impressed him most about DeepSeek’s success was not the increased danger from Chinese competition but the importance of keeping AI models open source so that everyone could benefit.
LeCun said it’s not that China’s AI is better than the US’s; it’s that open source models are outperforming those that are owned by companies.
“DeepSeek has benefited from open research and open-source tools like PyTorch and Llama from Meta. They developed new ideas by building on the work of others,” LeCun added. “Since their work is published and freely available, anyone can benefit from it.” This shows the strength of open study and open source.
Shortly after DeepSeek’s newest version came out, Meta CEO Mark Zuckerberg said the company plans to invest more than $60 billion in 2025 to continue focusing on AI.
Meta’s new strategy is designed to strengthen its standing against competitors OpenAI and Google in the competition for AI leadership.
Big tech companies have spent a lot of money, tens of billions of dollars, to build AI systems since OpenAI’s ChatGPT became popular.
Meta’s announcement came just days after Trump announced that OpenAI, SoftBank and Oracle will join a venture called Stargate and invest $500 billion in AI infrastructure across the U.S.
Marc Andreessen, a Silicon Valley investor who has been coaching Trump, wrote on X that “Deepseek R1 is one of the most impressive breakthroughs I’ve ever seen.” “It is open source, which is a great gift to everyone.”
“DeepSeek R1 is AI’s Sputnik moment,” Andreessen wrote in a reply.
Alexandr Wang, the CEO of Scale AI, a software business in San Francisco, commented on the technology and said that DeepSeek’s rapid success is a “wake-up call for America.”
“DeepSeek is a wake-up call for America, but it doesn’t change the plan,” Wang said in a post on X.
Wang said that the USA needs to be more innovative and move faster, like it has throughout the history of AI. He also mentioned that we should strengthen rules on exporting chips to keep our advantage in the future.
“Wang said that all major advancements in AI have come from America.”
DeepSeek was developed by Liang Wenfeng, a Chinese hedge-fund manager who is now a key figure in China’s efforts in artificial intelligence, according to the Journal.
Experts told the Journal that DeepSeek’s technology is not as advanced as OpenAI and Google. It is a strong competitor even though it uses simpler and less advanced chips, and sometimes misses steps that U.S. makers think are important.
On Saturday, the Journal reported that two DeepSeek models were listed in the top 10 on Chatbot Arena, a site created by researchers at the University of California, Berkeley, that rates how well chatbots work.
Also Read: Rednote in China: What You Need to Know About the App a Lot of Tiktok Fans Are
DeepSeek’s main model is free, but the Journal said the company charges users who link their own apps to DeepSeek’s model and computer resources.
What do you say about this story? Visit Parhlo World For more.