The Blubber

Chinese AI platform DeepSeek has overtaken ChatGPT on Apple's, AAPL, downloads rankings

DeepSeek's R1 Outperforms U.S. AI Models, Sparking Open-Source Debate

Chinese AI company DeepSeek has made waves in the tech world with its latest R1 model, which has outperformed leading American AI firms such as OpenAI, Meta, and Anthropic in third-party benchmarks.

LeCun: Open-Source AI is Winning, Not Just China

For Meta’s chief AI scientist Yann LeCun, the biggest takeaway from DeepSeek’s success was not the rising threat of Chinese competition—but rather the power of open-source AI.

“It’s not that China’s AI is surpassing the U.S., but rather that open-source models are surpassing proprietary ones,” LeCun posted on Threads.

Both DeepSeek’s R1 and Meta’s Llama are open-source models, allowing researchers and developers worldwide to modify and improve them. In contrast, OpenAI, which was originally founded with a mission to make AI open and accessible to all, has shifted toward a closed-source approach.

LeCun emphasized that DeepSeek "profited from open research and open source", building upon innovations from the global AI community.

"They came up with new ideas and built them on top of other people's work. Because their work is published and open source, everyone can profit from it," he wrote. "That is the power of open research and open source."

DeepSeek’s R1 Shocks Silicon Valley

When DeepSeek unveiled R1 on January 20, the company claimed it demonstrated "remarkable reasoning capabilities" and was "pushing the boundaries" of open-source AI. The launch took Silicon Valley by surprise, dominating tech discussions during a week filled with the World Economic Forum, TikTok controversy, and President Donald Trump’s first days back in office.

Days after the announcement, Meta CEO Mark Zuckerberg revealed that his company planned to spend over $60 billion in 2025, doubling down on AI investments. Zuckerberg has long championed open-source AI and has described building open platforms as a key priority for the next generation of technology.

“Part of my goal for the next 10-15 years is to build the next generation of open platforms and have the open platforms win,” Zuckerberg said in September. “I think that’s going to lead to a much more vibrant tech industry.”

Open-Source vs. Closed-Source: A Growing Divide

Supporters of open-source AI argue that it accelerates innovation and democratizes access, allowing anyone to modify and improve the technology. Critics, however, claim that closed-source models offer better security and control, preventing misuse.

OpenAI CEO Sam Altman has defended his company’s closed-source approach, arguing that it provides "an easier way to hit the safety threshold." However, in a Reddit AMA last November, he suggested that OpenAI “would like to open-source more in the future.”

With DeepSeek’s R1 shaking up the industry, the debate over open vs. closed AI is heating up—potentially shaping the future of artificial intelligence development worldwide.