It took only 65 days for 5 people plus AI to complete the workload that 50 people would finish in a year. Right after laying off 16,000 employees earlier this year, Jeff Bezos claims: AI won't "take away jobs" — instead, there will be a "labor shortage" in the future?

If a company's CEO tells you that a project that used to take 50 engineers a full year to complete has now been successfully delivered by just 5 people in 65 days, what would be your first reaction?

Would it be marveling at the productivity revolution brought by AI, or feeling anxious for the other 45 employees?

Recently, Amazon CEO Andy Jassy shared an internal case that once again pushed the topic of AI and employment into the spotlight: He revealed that the company recently completed a large-scale internal system refactoring — with the help of AI, only 5 employees finished the work that originally required 40 to 50 people and about a year to complete.

Paradoxically, just a month ago, Amazon founder Jeff Bezos publicly stated at the VivaTech conference held in Paris, France: AI will not only not cause humans to lose jobs, but may even lead to labor shortages in the future.

01

5 employees, 65 days, completing the work that used to take 50 people a year

The story originated from a recent public sharing by Andy Jassy.

He introduced that Amazon once had a large-scale system refactoring project. According to past experience, this work usually required a team of 40 to 50 engineers and lasted about 12 months to complete. But this time, the company formed a team of only 5 members who were proficient in AI tools, and finally completed all the deliveries in just 65 days.

In terms of time, the project cycle was compressed from one year to more than two months; in terms of personnel scale, it only used about one-tenth of the manpower of the past — this is not a proof of concept in the laboratory, but a real production practice that took place inside Amazon.

But at the same time, a more realistic question also emerged: If 50 people were needed in the past and only 5 are needed now, what about the rest?

This concern is not unfounded. As early as January this year, Amazon announced that it would cut about 16,000 jobs worldwide.

At that time, Amazon's explanation was that it hoped to achieve three goals through organizational restructuring:

• Reduce management layers;
• Improve teams' sense of ownership and autonomous decision-making capabilities;
• Streamline processes and reduce internal organizational bureaucracy.

For US employees, Amazon provided a 90-day internal transfer window, allowing laid-off employees to apply for other positions.

In the internal letter issued by Beth Galetti, Senior Vice President of Human Resources, Amazon did not even define this as an "AI layoff", but emphasized that it was just an organizational optimization, and stated that the company would continue to recruit and continuously invest in the most important strategic directions in the future.

However, almost everyone knows that the so-called "most important strategic direction in the future" of Amazon is actually AI.

02

AI investment has entered the "money-burning" stage

In fact, over the past year, Amazon has almost bet all its growth focus on AI. According to Amazon's latest first-quarter financial report:

AWS cloud computing business revenue reached 37.587 billion US dollars, a year-on-year increase of 28%;

Capital expenditure for a single quarter was as high as 44.203 billion US dollars;

The company expects full-year capital expenditures to reach about 200 billion US dollars, a record high.

Most of these huge sums of money will be used for AI infrastructure construction, including data centers, GPU clusters, self-developed Trainium chips, and generative AI services.

In response, the capital market gave a positive response: For investors, AI means higher profit margins, lower labor costs, and faster product delivery speeds; but for ordinary employees, this is obviously not a positive signal.

In an earlier interview, Andy Jassy made it clear that with the popularization of generative AI, the number of white-collar positions within Amazon in the next few years will gradually decrease. From the perspective of enterprise management, this is called "improving efficiency"; but for those who receive layoff notices, there is basically no difference between the two.

03

But Bezos said: AI will eventually lead to a "labor shortage"

While the outside world is constantly discussing whether AI will take away jobs, Bezos has put forward a completely different judgment.

On June 17 this year, at the VivaTech conference in Paris, someone asked Bezos: Will AI make human labor worthless? Bezos responded: "I completely disagree with this view. In fact, I think AI will lead to a labor shortage."

In his view, AI will greatly lower the threshold for product development, entrepreneurship, and innovation: In the past, a startup team might need dozens of engineers, but in the future, with the help of AI, a few people or even one person can complete a lot of work. Therefore, more and more people will try to start businesses, develop software, and build new businesses, which may also give rise to new professions that do not exist today.

Therefore, he believes that what is truly scarce in the future is not job opportunities, but people who can seize these opportunities.

Looking back at history, this logic is not unfamiliar: The First Industrial Revolution gave birth to factory workers; The Internet has created new professions such as programmers, e-commerce operators, and content creators. Every technological revolution will eliminate some positions while creating new demands.

So, Bezos believes that AI will also follow this historical pattern.

The problem, however, is that many research institutions believe that generative AI is not completely the same as past automation: Previously, industrial automation mostly replaced repetitive physical labor; but now, large models are beginning to enter knowledge-intensive positions that were once considered relatively safe, such as software development, UI design, customer support, office document processing, marketing, and even content creation.

Goldman Sachs, an international investment bank, previously estimated that AI currently causes about 16,000 jobs to disappear in the United States on average every month, and the groups most affected are precisely young people, fresh graduates, and junior employees — in this regard, the recruitment trends of many enterprises in the past year have also confirmed this change: More and more companies are beginning to reduce the recruitment of junior engineers, and invest more budgets in AI tools and senior talents, hoping to improve individual combat capabilities with the help of AI.

In other words, AI does not completely replace senior engineers, but first compresses a large number of basic tasks required to train new employees.

04

The real contradiction may lie in the "time lag"

Many research institutions have actually made predictions about whether AI will create more jobs, and the views of Bezos and the market are not completely contradictory.

For example, the market research institution Gartner once predicted that starting from 2028, the number of new jobs created by AI is expected to exceed the number of positions replaced. If this prediction comes true, then in the long run, AI may indeed bring new industries and more employment opportunities just like the Internet.

But even if Bezos is talking about the end point several years later, many ordinary employees are facing the immediate changes: For a person who loses their job today, the new positions that will appear in three years cannot solve the current rent, loan, and living pressure. In the AI era, skill upgrading and career transformation have almost become compulsory courses, but not everyone has sufficient time, funds, and learning resources to complete this transformation.

So, what is your answer to the question "Will AI create more jobs?"

Reference link: https://www.ibtimes.co.uk/amazons-ai-revolution-fewer-workers-faster-results-1808628

This article is from the WeChat official account "CSDN", author: ZHENG Liyuan, published with authorization from 36Kr.