MIT Just Found The Cause Of The AI Bubble

Headlines in the last two years suggest programmers are being laid off, junior developers struggle to find work, and tech companies are freezing hiring because one AI tool can now do the work of a dozen entry-level engineers.

The presenter states that AI more often replaces tasks within jobs rather than entire jobs.

Current economic systems, including GDP, unemployment figures, and wage data, are built around counting jobs and people, not analyzing tasks within jobs that AI can perform.

According to the MIT Iceberg Index study, AI can technically perform work accounting for approximately 2.2% of the total US labor market wage value, which is about $211 billion, within the tech sector.

The MIT Iceberg Index study found that when applied to the entire US economy, the wage value of work AI can technically perform jumps to approximately 11.7% of the total US labor market, or about $1.2 trillion.

The MIT team developed the Iceberg Index by creating a digital representation of 151 million American workers across 923 occupations and 3,000 counties.

The MIT team cataloged more than 13,000 production-ready AI tools and applied the ONET skill taxonomy to them to profile their capabilities.

The Iceberg Index measures how much of the wage value inside each job AI can technically perform, rather than predicting job losses.

More than 100,000 job losses in 2025 were linked to AI restructuring.

Entry-level employment in AI-exposed occupations has decreased by 14% compared to the pre-ChatGPT era.

In 1965, Princeton economists William Baumol and William Bowen noted that a string quartet performing Beethoven required the same number of musicians and time in the 19th century as it did a century later, even though the cost of the performance had risen.

The presenter predicts that if the MIT Iceberg Index is not adopted, a two-speed economy will emerge, characterized by one half becoming dramatically more productive and the other half steadily less affordable.