Previously, I wrote an article about how American computer science graduates are struggling to find jobs. Some online commenters suggested that programmers would never become unemployed even if others did. I understand this sentiment, as programmers still seem to be in high demand domestically. However, the situation in America is quite different.

Data shows that enrollment in Duke University's introductory computer science courses plummeted by 20% this year. Classrooms that were packed five years ago now have plenty of empty seats. In 2025, the unemployment rate for American computer science students soared to 5.7%, while the overall national unemployment rate is only 4%.

Recently, Ford CEO Jim Farley stated that "artificial intelligence will replace half of America's white-collar workers." This isn't just talk—American companies are actually implementing this strategy. In July this year, Anthropic released a new AI platform promising to handle entire finance teams' work, from market analysis to investment memos. CEOs at companies like Shopify and Duolingo have mandated that hiring is only permitted if it can be proven that AI cannot perform the task.

Consequently, many Americans are transitioning to blue-collar jobs. This aligns with AI pioneer Geoffrey Hinton's prediction last year that "the safest profession in the future might be plumbing," since AI struggles with such hands-on work.

AI has arrived, but its distribution is uneven. Why are American white-collar workers the first to be impacted?

The answer lies in America's smaller population, higher labor costs, and advanced technology. The country has always preferred using software to replace human labor. Simply put, America's industrial internet development is far ahead of China's.

Industrial internet refers to enterprise service companies using software to help businesses undergo digital transformation and reduce costs while improving efficiency. As a ten-year veteran in the industrial internet sector, I understand the difficulty of finding clients (many deals rely on relationships). For most companies, digitization software is a luxury—they prefer using human labor.

A developed industrial internet means America focuses more on specialized software, pursuing technological excellence and leading globally in the market. This reflects America's strong tool-oriented thinking and its many globally dominant software giants.

As a new tool, AI naturally disrupts old tools first. Take design industry leader Adobe, for example. Previously, photo editing and design work using Photoshop took hours. Now, ordinary users can input a few sentences and AI tools like Midjourney can produce professional images in seconds. In December 2024, Adobe's stock price experienced its largest decline in two years.

Consider Salesforce.com, the world's largest customer relationship management (CRM) software company. Previously, employees had to manually analyze customers and write emails. Now AI completes these tasks in seconds. This year, Salesforce's core module growth has significantly declined.

Alphabet's search business has been particularly hard hit by AI disruption. This year, its global market share fell below 90% for the first time (according to StatCounter data). Search results with AI summaries saw click-through rates plummet from 15% to 8% (Pew Research Center), and user website click rates dropped by nearly half.

Why has China experienced much less AI impact? Take Baidu, which also started with search. Baidu doesn't face Google's concerns because China's consumer internet is too developed and competitive. Baidu's search business was disrupted long ago by WeChat, Xiaohongshu, and TikTok. People have been using various social media platforms to search for information—reading official accounts for news and searching Xiaohongshu for lifestyle content.

Baidu transformed into an AI company early on. Today's second-quarter earnings report shows Baidu's AI new business revenue exceeded 10 billion yuan, growing 34% year-over-year. Alphabet's challenge is that it has relied on search revenue for too long and earned global profits for so long that it's now struggling like a turning elephant when facing AI disruption.

It's not just Alphabet. In consumer internet, Chinese giants clearly have the upper hand. Musk has long envied WeChat's integration of social media and payments. ByteDance's TikTok competes globally in social media and continues rising through social commerce combinations.

But does AI significantly impact China's consumer internet? It's more about empowerment than revolution.

AI impacts American giants more because American tech companies excel at creating "tools"—like Adobe's design software and Salesforce's enterprise software. As a new tool, AI fundamentally revolutionizes these old tools, directly threatening their "golden rice bowls."

Chinese internet giants are better at "traffic" businesses, like WeChat and Taobao. For them, AI is more like a new weapon that improves efficiency, helping consolidate and optimize existing traffic models rather than fundamentally disrupting them.

Of course, while American giants face greater impact, they operate globally, have more money, and invest more aggressively in AI. According to the latest earnings forecasts, in 2025, Alphabet, Microsoft, Amazon, and Meta's combined capital expenditures are expected to exceed $400 billion, mainly for AI infrastructure including chips and data center computing power. This figure nearly equals the annual GDP of many moderately developed countries.

Chinese internet giants also invest money, but in much smaller amounts. According to relevant reports, in 2025, China's "Big Four"—Alibaba, Baidu, Tencent, and ByteDance—are expected to have combined capital expenditures of $51 billion, much less than American giants, mostly defending rather than aggressively attacking.

Beyond enterprises, the Chinese and American governments have completely different attitudes toward AI—one aggressive, one moderate. Last month, America launched an artificial intelligence action plan, clearly emphasizing development over regulation. Trump knows that to make America great again, he must win the AI race.

China is obviously more stability-focused, emphasizing macro advancement and secure development.

The current situation shows America leading in closed-source technology while China pursues open-source ecosystem building to catch up. Having recently finished reading "Fire in the Valley," I'm struck by how computer enthusiasts' selfless sharing once enabled computer development, and the internet's birth resulted from the World Wide Web being open source. Now history has quietly reversed—America's OpenAI isn't "open," while China chooses open source to build larger ecosystems in its pursuit to catch up.

We can hardly predict the final outcome, but one thing is certain: first place won't always remain first place.