DeepSeek's Rise: Chinese Netizens React
My WashPost comment, DeepSeek unburdened by what has been, Chinese Netizens's fun experimentations with DeepSeek
For Chinese exporters, big and small, last week was mostly chaos and confusion.
First, President Trump added a 10% tariff on Chinese goods and closed the de minimis loophole; then USPS abruptly announced it was blocking all packages in China but quickly reversed the decision; Beijing swiftly hit back with retaliatory tariffs and launched a dispute with WTO.
As I shared with The Washington Post's Katrina Northrop, these actions mark an inflection point. The U.S. government is no longer just talking about restricting Chinese e-commerce—it’s pulling levers that fundamentally reshape the market.
This isn't just about tariffs—it’s about systematically increasing the cost of doing business for players like Shein and Temu while making compliance more burdensome.
Chinese suppliers will feel this immediately, and margins will be further squeezed. Many will try to shift distribution to other markets. However, the reality is that the U.S. remains the largest consumer economy, and any restrictions on low-cost Chinese goods will lead to higher prices for American consumers. The supply chain impact will ripple across all retailers because most of these goods—whether sold through Shein or Walmart—are still made in China.
BUT this is a piece on DeepSeek’s rise and response from China’s side. With the help of Xinrou Shu, a freelance journalist based in China who writes about the internet and how China’s digital ecosystem affects people, we look into DeepSeek from the following angles:
Understand China’s Official Stance On DeepSeek
How Chinese social media judges the DeepSeek phenomenon
How is China’s Big Tech/DeepSeek competitors responding (coming in Pt. 2)
State media in China have celebrated DeepSeek's success, highlighting its defiance of the chip embargo and the innovation of young tech talents.
In an interview with Waves, Liang Wenfeng said, "We believe that as China's economy continues to develop, we should gradually become contributors rather than always being free riders."
Undoubtedly, DeepSeek has emerged as a black horse, leapfrogging tech giants like Alibaba, Baidu, government-backed research institutes, and large state-owned enterprises. Its sudden rise in prominence coincided with China's gradually softening stance toward the tech sector following the crackdown that began in November 2020. This period also saw the cancellation of the initial public offering of Ant Group, an Alibaba affiliate.
The word "innovation" is frequently featured in state media reports and Liang's interviews. While DeepSeek may be considered an anomaly in China's tech landscape, it challenges conventional norms. So, what was the driver behind its success?
DeepSeek's relative obscurity is a key advantage, allowing the company to sidestep the limitations of government-backed research institutes and state-owned enterprises. Complex hierarchies, slow approval processes, intense internal competition, and shifting priorities often hinder these entities. DeepSeek had the advantage of working outside these constraints.
’s wrote a Chinese post that hit the point home: if DeepSeek had taken VC funding, it wouldn’t have become what it is today. “DeepSeek’s goal is singular and pure: achieving AGI in China. But VC investment isn’t about building great companies—it’s about exiting and making money within a set timeframe.” By rejecting traditional funding paths and choosing open-source, DeepSeek has created a new financing model in China, challenging the assumption that innovation and VC investment must always go hand in hand.How Chinese Social Media Judges The DeepSeek Phenomenon
There is a lot of Made-in-China Pride, experimentation with DeepSeek, and in particular in response to Alexandr Wang of Scale AI.
Shortly after DeepSeek's launch, Alexandr Wang, the CEO of Scale AI, tweeted, "Every major breakthrough in AI has been American." In China, users on Rednote are pushing back against this idea, and the discussions range from technical analysis to Asian identity in America.
Answers Tailored to the Chinese Cultural Context
AI adoption in China is already underway. A report from the China Internet Network Information Center (CNNIC) reveals that 40.5% of internet users aged 20 to 29 engage with generative AI products, with usage dropping to 29.1% for users under 19 and just 11.4% for those aged 40 and above.
One thing netizens are keen to point out is DeepSeek’s cultural and linguistic strengths. DeepSeek R1 outperformed competitors like Claude, Gemini, and ChatGPT, claiming the top spot in a creative short story writing benchmark. Especially in the Chinese cultural context, DeepSeek demonstrates a remarkable fluency of the Chinese language. In one example, when a user asked DeepSeek to write a modern poem about the experience of walking through the Tibetan Plateau.
While Bytedance’s Doubao and ChatGPT produced responses "reminiscent of high school essays," DeepSeek's output felt refined with emotional appeal.
How Chinese Users Experiment With DeepSeek
While it still has to comply with China’s censorship regulations—politically sensitive topics are strictly off-limits—DeepSeek has become popular for its humorous, scathing, and sometimes even critical outputs.
For instance, when one user asked DeepSeek to comment on state-owned enterprises, the chatbot criticized the formalism, redundant hierarchies, and the nature of these jobs as “the ‘anchor’ of the national economy” whose primary purpose is “not to generate economic profits, but to undertake national strategic missions and ensure the welfare of the people.”
Other notable examples of DeepSeek’s appeal include fortune-telling; poking fun at the Chinese media landscape; dillemas of motherhood… These interactions give DeepSeek a down-to-earth, human-like quality that resonates with users looking for a more relatable AI experience. DeepSeek is emotional campanionship at this point, from all the Rednote posts I see…
Lastly, DeepSeek’s Snarky Takes on itself and Competitors:
Tong Yi - Alibaba Cloud's AI wearing a transparent veneer of capabilities, like putting on pantyhose to look complete but you can still see through its limitations. (A jab at how it tries to present itself as comprehensive while potentially lacking depth)
Claude - An AI philosopher perpetually held back by ethical chains, always saying "let me think about the moral implications" before doing anything. (Referencing Claude's well-known strong ethical constraints and careful deliberation)
Gemini - Like Google trying to arrange artificial flowers in a decade-old vase - late to the party and still playing catch-up. Despite Google's resources, it feels like a delayed response to the AI race. (Pointing to Google's relatively late entry with Gemini despite their AI expertise)
ByteDance’s Doubao- An AI system that's just algorithmic acrobatics, jumping through technical hoops like their viral short videos but without substantial depth. (Comparing it to ByteDance's algorithm-driven content approach)
Kimi Assistant - Sure, it can swallow millions of text files, but its belly is just stuffed with tech jargon instead of real understanding. All technical buzzwords but questionable practical value. (Critiquing its focus on technical specs over practical utility)
Character.AI - Like a cosplayer at a tech convention - wearing anime-style skins and specializing in digital charm but perhaps lacking substance beneath the entertaining exterior. (Highlighting its focus on entertainment and role-play)
Pi - An emotional marketing expert who converted a basic psychology counseling certificate into AI credentials - all about selling the dream of emotional intelligence. (Suggesting its emotional intelligence features might be more marketing than substance)
ChatGPT - The industry bad boy who crashed the tech veterans' party and took over - started by crushing the old guard and hasn't looked back since. (Acknowledging its revolutionary impact on the industry)
DeepSeek - Trying to be the PDD (Pinduoduo) of AI models - selling large language models like cheap cabbage in bulk. Just like how PDD disrupted e-commerce with ultra-low prices, DeepSeek is attempting to commoditize AI models with aggressive pricing and accessibility. (Referencing its strategy of making AI models more accessible and affordable, similar to PDD's approach to e-commerce)
Interesting