I built this because I was frustrated by the AI-and-jobs conversation. Every take is either dystopian or dismissive, and neither side cites systematic data.
The IDI aggregates 19 public signals into a single 0-100 score across four equally weighted categories: Job Displacement, AI Capacity & Cost, AI Penetration, and Physical Automation.
Current score: 43 (Moderate).
All sources are free/public (BLS, FRED, Indeed, METR, arXiv, CA DMV, Anthropic Economic Index, Amazon filings). 6-month EMA smoothing, domain-specific normalization. 85 months of history, downloadable as JSON (CC BY 4.0).
A few things that surprised me building this:
The biggest signal isn't job losses. It's the collapse in model pricing: $120/MTok to $0.20/MTok in under two years. METR's agent horizon went from 3.5 minutes to ~6.5 hours in the same period. Yet Job Displacement only scores 45. There's a growing gap between what AI can do and what it's actually doing to labor markets.
Counter-signals are real. New business formations are up 23% from 2022. Real wages in tech are still growing. The honest version of this story includes adaptation, not just disruption.
Weighting and signal selection are editorial calls, documented on the page.
Happy to discuss methodology or take suggestions for signals I'm missing.
We just built a simple tool to evaluate if your cold email will generate replies.
You can also submit it to get a detailed report with advices to improve it.
Data comes from analyzing reply rates from our customers and confirmed by the following research paper : "Characterizing and Predicting Enterprise Email Reply Behavior" (Yang et al. 2017)
Let me know what you think about it and how I could improve it :)
I built this because I was frustrated by the AI-and-jobs conversation. Every take is either dystopian or dismissive, and neither side cites systematic data.
The IDI aggregates 19 public signals into a single 0-100 score across four equally weighted categories: Job Displacement, AI Capacity & Cost, AI Penetration, and Physical Automation.
Current score: 43 (Moderate).
All sources are free/public (BLS, FRED, Indeed, METR, arXiv, CA DMV, Anthropic Economic Index, Amazon filings). 6-month EMA smoothing, domain-specific normalization. 85 months of history, downloadable as JSON (CC BY 4.0).
A few things that surprised me building this:
The biggest signal isn't job losses. It's the collapse in model pricing: $120/MTok to $0.20/MTok in under two years. METR's agent horizon went from 3.5 minutes to ~6.5 hours in the same period. Yet Job Displacement only scores 45. There's a growing gap between what AI can do and what it's actually doing to labor markets.
Counter-signals are real. New business formations are up 23% from 2022. Real wages in tech are still growing. The honest version of this story includes adaptation, not just disruption.
Weighting and signal selection are editorial calls, documented on the page.
Happy to discuss methodology or take suggestions for signals I'm missing.