A person uses DeepSeek app on a mobile phone on Feb 17, 2025. [Photo/Xinhua]
Recently, when I realized that even my 60-year-old father was using DeepSeek, I knew something had changed fundamentally.
My dad, a cautious man by nature, is no fan of digital trends. He rarely watches short videos and has never been intrigued enough to download Douyin or Kuaishou. His time online is mostly spent browsing news and current affairs — no flashy distractions, just information.
So, when I found him enthusiastically tapping away on his phone one evening, asking DeepSeek for the latest market updates, I paused.
In China, DeepSeek is somehow finding its way into the most unlikely hands — from healthcare to capital markets to daily routines, it has found its niche everywhere.
People are using it to search for information, ask questions, chat, organize work — and, believe it or not, even tell fortunes.
That was the moment I realized the world had changed. Technology, once distant and complicated, has become something familiar, something human. And behind this technology, China is closing its gap with the United States in terms of AI development.
The 2025 Stanford HAI Report on AI, released recently, is packed with statistics and projections about global AI trends, but one headline stands out: The performance gap between top Chinese and US AI models has narrowed considerably.
The world’s balance of technological power is shifting, and DeepSeek is partly the reason.
Two years ago, I remember writing a story asking “When will China have its own version of ChatGPT?”
That seemed like a distant dream, an ambition that would take years to materialize. But in 2025, that dream has become a reality.
According to the Stanford University report, the performance gap between top Chinese and US models, which had been a staggering 20 percent in 2023, has shrunk to nearly nothing — 0.3 percent in 2025.
Open-weight models, like Deep-Seek, have made jaw-dropping progress.
The gap between open-weight and closed-source giants, which was 8 percent in 2024, has dropped to just 1.7 percent in 2025. In the tech world, those numbers are nothing short of revolutionary.
Of course, the US still holds the lead when it comes to the big names in AI.
In 2024, 90 percent of the world’s most well-known models had been developed by commercial firms, with the US contributing 40 of them and China only 15. But what is undeniable now is the fact that the gap between the best models is evaporating.
In 2024, the difference between the best and the 10th-best model was 12 percent. Today, that difference has shrunk to just 5 percent, signaling a convergence in AI capabilities. We are no longer trailing behind.
And as if that isn’t enough, Deep-Seek’s growth trajectory has been staggering. The numbers don’t lie.
According to analytics from aitools.xyz, DeepSeek has overtaken OpenAI’s ChatGPT in monthly website visits.
In February alone, it recorded 525 million visits, surpassing ChatGPT’s 500 million. For a product that had only launched its V3 large model a few months earlier, these numbers are mind-blowing.
By March, DeepSeek had already captured 6.58 percent of the global AI tools market. To put that in perspective, ChatGPT still led with 43 percent, but DeepSeek had already approached AI tool Canva (8 percent).
It is clear — this isn’t just a domestic phenomenon anymore. Deep-Seek has gone global, and it is only getting started.
Watching my father use DeepSeek made me realize that technology is no longer a foreign language. It is something we can all speak.
A silent revolution has taken root, quietly changing how we live, work, and interact with the world. And as I thought about those 525 million visits, I couldn’t help but wonder — just how many fathers are embarking on this AI journey?
Tanks to chinadaily.com.cn
Please visit:
Our Sponsor