Karpathy’s AI Job‑Risk Chart Sparks Panic: 60 Million White‑Collar Jobs Threatened
Karpathy’s quickly‑removed project scores 342 U.S. occupations for AI exposure, revealing an average risk of 4.9, with 42% of 60 million white‑collar jobs scoring 7 or higher, while physically‑oriented roles like plumbers remain safe; Harvard research corroborates these findings and shows a nuanced shift toward job augmentation rather than wholesale replacement.
Last night Andrej Karpathy published a fast‑built site (karpathy.ai/jobs) that evaluates AI‑replacement risk for 342 U.S. occupations using data from the Bureau of Labor Statistics. Each job received an exposure score from 0 to 10, and the average across all occupations is 4.9.
High‑Risk White‑Collar Jobs
The analysis shows that about 42% of jobs (≈60 million positions) have a risk score of 7 or higher. Notable high‑risk roles include software developers (9/10), medical transcriptionists (10/10), lawyers (8/10), and office clerks (9/10). These positions typically involve routine text processing, data entry, or standardized workflows, making them highly susceptible to automation.
For example, office clerks have a median salary of $43,630 and a job count of 2.6 million, while financial analysts earn a median of $101,910 with 429 k positions, yet both score above 8 on the exposure index.
Low‑Risk Physical Jobs
Jobs that require complex physical interaction—such as plumbers, pipefitters, roofers, and other manual trades—score between 1 and 3, indicating low AI exposure. A plumber with a high school diploma earns a median salary of $62,970 and is among the safest occupations.
Broader Academic Context
Harvard Business School researchers (Srinivasan et al.) released a working paper that tracks U.S. online job postings from 2019 to early 2025. Using GPT‑4o to evaluate 19,000 tasks across 900 occupations, they compute both automation and augmentation scores. Their findings align with Karpathy’s: high‑automation roles see a 17% quarterly drop in hiring, while high‑augmentation roles experience a 22% increase.
In high‑automation occupations, demand for AI‑related skills falls 24%, and overall skill requirements shrink. Conversely, in high‑augmentation roles, AI‑skill demand rises 15% and total skill breadth expands.
Implications
The combined evidence suggests AI is reshaping the labor market by eliminating routine, information‑processing tasks while elevating roles that require physical presence, complex judgment, or human‑centric interaction. Workers are urged to assess how much of their job can be automated and to upskill accordingly.
All referenced charts and data visualizations are retained as images for context.
Machine Learning Algorithms & Natural Language Processing
Focused on frontier AI technologies, empowering AI researchers' progress.
How this landed with the community
Was this worth your time?
0 Comments
Thoughtful readers leave field notes, pushback, and hard-won operational detail here.