AI will take jobs, and an era of unemployment is coming. I keep hearing stories like this, but honestly, I think it’s a complete misunderstanding.

A16z’s partner explains it in detail, but essentially the premise itself—“the total amount of work is fixed”—is wrong. Human needs and desires are constantly evolving, and the cheaper technology becomes, the more new demand is created.

Look at history. In the early 20th century, one-third of the American labor force was working in agriculture. But after tractors became widespread, agricultural workers didn’t end up unemployed; instead, they moved into factories, stores, offices, hospitals, and even the software industry. Agricultural output tripled, yet unemployment didn’t increase. The same was true for electrification. In the early 1900s, only 5% of American factories used electricity, but by 1930, 80% were operating with power. After that, new industries—manufacturing, sales, credit provision, and more—kept emerging one after another.

When VisiCalc and Excel appeared, people said bookkeeping jobs would be over. But in reality, bookkeeping decreased, while the number of financial analysts increased by 1.5 million. Every time productivity improved, the whole economy expanded.

The same applies to AI: what matters is what today’s data shows. Multiple academic studies indicate that AI adoption hasn’t brought about major changes in total employment. According to Goldman Sachs’ estimates, the productivity-enhancing effects of AI are far greater than the substitution effects. In management’s earnings call presentations too, the term “augmentation” is used about eight times as often as the term “substitution.”

Demand for software engineers is rising rapidly. Since early 2025, jobs in software development have increased steadily in both number and share, and the number of product manager job openings has reached an all-time high. That’s because AI is driving coding—new apps and businesses are exploding as a result.

Sure, routine administrative work and some customer service roles may decline. But at the same time, analysis, technical, and management work are being enhanced and complemented by AI. That’s the true role of pioneering technology: to rebuild economic structures and expand the scope of useful work.

Most jobs created after 1940 didn’t exist in 1940. People back then couldn’t have imagined job roles like today’s cloud engineers or AI specialists. The same thing is happening now. The emergence of new businesses is accelerating, and robotics-related datasets jumped from 10th place to 1st place in just 2 years.

At the macro level, there’s still no statistically significant relationship between AI and the unemployment rate. Some jobs disappear, some emerge, some lose value, and some gain value. In other words, the labor market is not static—it’s always evolving.

In conclusion, AI brings affordable intelligence. What that means is a larger market, new companies, new industries, and more advanced human work. There is no fixed amount of work, and no fixed amount of cognitive ability. It has been that way before, and it will be that way in the future. Doomsday talk is just a lack of imagination.