Karpathy’s Chilling Visualization Shows Which Jobs AI Could Replace

Karpathy’s new "karpathy/jobs" visualization scores 342 U.S. occupations on AI exposure using Gemini Flash, revealing that software engineers rank 8‑9, transcriptionists hit 10, while roofers stay near 0, highlighting which jobs may soon be automated.

AI Engineering
AI Engineering
AI Engineering
Karpathy’s Chilling Visualization Shows Which Jobs AI Could Replace

Project Overview

karpathy/jobs is an interactive visualization that quantifies AI‑replacement risk for every occupation listed in the U.S. Bureau of Labor Statistics (BLS) Occupational Outlook Handbook.

Data Collection and Scoring Method

All 342 occupations from the BLS dataset were scraped.

Each occupation description was submitted to the Gemini Flash large language model.

The model returned a numeric exposure score from 0 (no exposure) to 10 (maximum exposure).

Scores were aggregated with a weight proportional to total employment to compute a weighted average.

The resulting scores are displayed as a hierarchical tree diagram.

Key Results

Weighted average score: 4.9 / 10.

Software engineer: 8‑9 / 10, indicating high exposure despite the technical nature of the role.

Transcriptionist: 10 / 10, the highest possible exposure.

Roofer: 0‑1 / 10, indicating current safety from AI automation.

Open‑source Artifacts

All code and data are publicly available at https://karpathy.ai/jobs, with the underlying BLS data taken directly from the Occupational Outlook Handbook.

Related Context

The findings align with a long‑standing observation that occupations whose core activity is “sitting in front of a screen processing information” are prone to automation. For additional perspective, see Anthropic’s AI‑employment impact report (https://mp.weixin.qq.com/s?__biz=MzA5MTIxNTY4MQ==∣=2461158737&idx=1&sn=2317b702b63165021280e46e8d2155f9&scene=21#wechat_redirect).

AI automationoccupational analysisKarpathyBLS dataGemini Flashjob risk
AI Engineering
Written by

AI Engineering

Focused on cutting‑edge product and technology information and practical experience sharing in the AI field (large models, MLOps/LLMOps, AI application development, AI infrastructure).

0 followers
Reader feedback

How this landed with the community

Sign in to like

Rate this article

Was this worth your time?

Sign in to rate
Discussion

0 Comments

Thoughtful readers leave field notes, pushback, and hard-won operational detail here.