Why I say "AI taking away human jobs on a large scale" is a fantasy

📄Full Article· Automatically extracted by trafilaturaGemini 翻譯2236 words

This article is based on an X post by David George, a partner at the a16z Growth Fund. Using 18 data charts, he argues that the narrative of AI causing mass unemployment is a fantasy. Throughout history, every major technological revolution (electricity, automobiles, spreadsheets, the internet) has destroyed specific jobs in the short term, only to create far more new jobs with higher wages in the long run. (Context: DeepL, the biggest rival to Google Translate, lays off 25% of staff! CEO: Transforming into an "AI-native" company, human employees will focus only on creative work in the future.) (Background: Cloudflare stock plunges 17% after-hours following "1,100 layoffs"; CEO: Shifting to an AI-first development model.) When a technology significantly lowers the production cost of a certain type of good, the size of that market explodes—the resulting demand and jobs often exceed the number of those originally replaced. In the 1920s, the manufacturing costs of durable goods (home appliances) plummeted in just ten years: coffee makers -69%, electric blankets -58%, radios -54%, fans -52%, and washing machines -44%. The result of this cost collapse was not factory closures, but the transformation of these products from elite items into mass consumer goods, driving durable goods manufacturers to outperform the broader market between 1924 and 1935. The story of the automobile is even more intuitive: in 1900, a new car cost about $45,000 (in 2024 currency), falling to under $5,000 by 1925. The result? Annual production in the U.S. surged from near zero in 1900 to over 3.5 million vehicles by 1925; employment in the automotive industry also grew from almost zero to over 400,000 people. Today, the "price of intelligence" for LLMs is falling at a logarithmic rate—from September 2023 to the end of 2025, the median cost has dropped by two orders of magnitude. This is precisely the starting point for the next wave of employment growth in history. There is often a gap of several decades between the advent of a technology and its actual impact on labor productivity. Faraday established the principles of electricity in the 1820s–1840s, Edison commercialized the light bulb in 1879, but it wasn't until the 1920s, when factories fully adopted unit drive motors, that labor productivity saw a structural leap—a gap of nearly a hundred years. The increase in agricultural productivity is another counterintuitive case: the real prices of corn, wheat, and rice have fallen by more than 80% since the 1910s, yet the world population has grown from 1.8 billion to 8 billion in the same period. Increased agricultural efficiency did not leave agricultural workers unemployed; instead, it fed more people and released labor to enter other industries. The internet nearly halved the profession of travel agents—the number of travel agents in the U.S. fell from a peak of 150,000 in 2000 to about 50,000 today. This is a real case where technology did indeed replace jobs. However, the wages of the remaining travel agents have continued to rise relative to the overall average. The ratio of travel agent wages to the overall average wage climbed steadily from 0.80 in 1990 to over 0.95 in 2025. Productivity tools have made the value of the remaining high-skilled workers even higher. The story of the spreadsheet is similar: after spreadsheets became widespread in the 1980s, the number of bookkeepers continued to decline from 2.1 million to under 1 million today; however, in the same period, the number of accountants and auditors grew from 650,000 to 2.1 million, and financial analysts grew from near zero to 1.5 million. Spreadsheets eliminated low-skilled clerical work but gave birth to the entire FP&A industry. Critics predicted that AI would make software engineers unemployed, but the data shows the opposite. The number of software engineer job openings on Indeed showed a clear reversal in mid-2025 and rebounded rapidly in early 2026, far exceeding the recovery speed of overall job openings. Not only has the total volume of job openings rebounded, but the share of software developers among all new jobs has also expanded, rising from about 1.6% in 2023 to over 2.2% in early 20

Data Status✓ Full text extractedRead Original (動區 BlockTempo)

🔍Historical Similar Events· Keyword + Asset Matching0 items

No similar events found (requires more data samples or embedding search; currently MVP keyword matching)

Raw Information

ID:ce5e36f356

Source:動區 BlockTempo

Published:2026-05-10 02:47:27

Category:zh_news · Export Category zh

Symbols:Unspecified

Community Votes:+0 / −0 · ⭐ 0 Important · 💬 0 Comments