I found a rather interesting debate—it’s a long-form article written by an a16z partner. It begins with the question: Is it really true that AI will end employment?

To put it simply, the article argues that the “doomsday” idea that AI will destroy jobs is, from an economic standpoint, wrong—and the reasoning is fairly convincing.

First, there’s a basic misunderstanding: the premise that “the amount of work is fixed.” They call this the “fallacy of total workload,” but it’s completely mistaken. Human needs and desires are not static, and when technology gets cheaper, demand tends to expand rather than shrink.

You can see this clearly from history. During the age of agricultural mechanization, one-third of the U.S. working population was involved in agriculture. This was in the early 20th century. By 2017, it had fallen to 2%. Did tractors take away jobs? It might look that way. But in reality, agricultural production increased threefold, and the unemployed agricultural workers flowed into factories, stores, offices, hospitals, research institutes, and eventually into service industries and the software industry. In other words, technology didn’t end up dominating existing jobs; instead, it rebuilt the economy as a whole and created new industries.

The same pattern holds for electrification. In the early 20th century, only about 5% of U.S. factories were using electricity. By 1930, about 80% were operating on power. In the following decades, labor productivity doubled. Yet unemployment didn’t rise. Instead, new industries kept emerging one after another.

When Excel was introduced, people said bookkeeping jobs would be over. Did that really happen? No. Bookkeeping clerks decreased by 1 million, but financial analysts increased by 1.5 million. So jobs didn’t disappear—they shifted to more advanced roles.

The example of travel agencies is also interesting. Since 2000, as technology has advanced, the number of travel agency employees has fallen by half. That’s true. But those employees didn’t become unemployed—they found work in other parts of the economy. And here’s what’s important: the wages of the employees who remained at travel agencies increased. In 2000, their pay was 87% of the all-industry average; by 2025, it had risen to 99%. In other words, when productivity increases, the value of the people who remain in that industry rises as well.

Something similar is happening with AI. According to a Goldman Sachs study, the “augmentation” effect from AI is far larger than the “substitution” effect. In earnings calls, the phrase “using AI as augmentation” was used about 8 times as often as “using AI as substitution.”

The growing demand for software engineers also proves this. Since AI is accelerating coding, the demand for developers is actually increasing. Since early 2025, jobs in software development have been steadily increasing in both headcount and their share of the overall employment market. The number of product manager job openings has also continued to rise and has reached the highest level since 2022.

So what do current data show? There are multiple academic studies. According to an NBER working paper, AI adoption hasn’t caused a major change in total employment, but it has been reshaping the distribution of tasks and occupations. Routine administrative work is easier to replace with AI, but analysis, technical, and managerial work are often described as being strengthened or complemented by AI.

In research from the Federal Reserve Bank of Atlanta, on average, more than 90% of firms estimate that “there has been no impact from AI in the past 3 years.” In research from the U.S. Census Bureau, only about 5% of companies using AI report any impact on employee headcount, and even among those, increases and decreases are roughly the same.

A study from Yale University concludes that AI’s impact on the labor market strongly reflects stability rather than major economic upheaval. That means, put another way, that AI has a neutral effect on employment.

There’s an important exception. Research by Stanford University and the Federal Reserve Bank of Dallas finds that it’s becoming harder to find entry-level jobs that are “deeply involved” with AI. But at the same time, they also find an increase in entry-level roles where AI plays a supportive role.

On a macro level, there is not yet a statistically significant correlation between AI and unemployment rates. Instead, “AI-augmented” industries tend to show higher employment growth rates and lower unemployment.

The emergence of new businesses is increasing explosively, and it has a strong correlation with AI applications. The share of new apps appearing in app stores increased by 60% year-over-year. The robotics field is also drawing attention: robotics-related datasets jumped from 10th place to 1st place in just 2 years.

Until AI truly changes society, there are many unfilled jobs in the robotics field. That means more jobs will be created going forward.

Of course, not every job is affected. The U.S. Bureau of Labor Statistics predicts that jobs such as customer service representatives and medical administrative assistants will decline. But that doesn’t mean overall unemployment will rise. Some jobs will disappear and some will shrink, but new jobs are created at the same time.

Most jobs created since 1940 did not exist in 1940. By 2000, the decline of travel agencies could be predicted, but it was difficult to imagine the emergence of a technology services industry for mid-sized companies centered on cloud migration—until cloud computing became widely adopted.

Knowledge work isn’t ending; it’s actually beginning. Automation eliminates repetitive tasks, raising human work to a higher level. Since humans crave expansion, once shortages disappear, people aim for further development at higher levels. When food prices fall, spending increases on housing, healthcare, education, travel, entertainment, convenience, pets, security, beauty, and longevity.

The same applies to the labor market. Human ambition never runs out, so new jobs keep being created. When you conquer an old frontier, new frontiers appear that need to be conquered again.

We shouldn’t view the modern economy as a museum of past jobs. Instead, it’s a creative system for allocating resources that continuously creates new jobs, new tasks, new goals, and new inventions.

From a macro perspective, the future isn’t an era of unemployment. Cheaper AI, larger markets, new companies, new industries, and more advanced human roles are ahead. There are no fixed quantities for the amount of work or for cognitive capacity. That has always been the case, and it will continue to be.

AI isn’t the end of work—it’s the beginning of a richer era of intelligence.