The Ledger

The Ledger: dispatches from the Approximation Era.

A running log of where AI is actually hitting the economy. I scan every earnings call and SEC filing, verify each deployment against independent sources, and publish what holds up. Newest first.
5,000+ verified, and counting Day 247 of documenting the Approximation Era.
4,198
Companies
6,103
AI Use Cases
1,376
Vendors
7
Sectors

Three things I found

3.9x Energy and Transportation produce nearly four AI winners for every underperformer. Finance and Retail are inverted. Physical-asset sectors dominate.
3.4% Of 1,376 AI vendors tracked, only 47 serve both high-performing and underperforming companies. The vendor market has split into two parallel ecosystems with almost no overlap.
+0.02 Companies with no reported AI activity carry a higher mean Information Ratio than those deploying AI. Most companies are doing AI badly. The quiet ones are outperforming.

Dispatches

Long-form entries in the log. Reverse chronological. Day numbers count from the first entry.
May 2026 Thesis · 9 pages
The Approximation Era
The math that brought John Glenn home is now writing the treatments for cancer. The thesis that ties Katherine Johnson's chalkboard to AlphaFold and the post-cancer future.
Read the essay →  ·  PDF ↓
April 25, 2026 Report · 8 pages
Where AI Creates Value — Q2 2026
5,752 AI deployments. 2,385 companies. Which deployment patterns correlate with market outperformance, and which five attributes separate the winners.
Download PDF ↓
April 16, 2026 Report · 10 pages
Industry AI Report — 1Q26
The thesis, methodology, and seven-sector summary. Anonymized. No company names.
Download PDF ↓
April 9, 2026 Note
Watts, Tokens & Money
Value flows downward in the AI economy. Energy wins, pure software loses. The companies that control watts and silicon will own the next decade.
Download PDF ↓
April 6, 2026 Note
Move to the Edge
AI is eating the middle of every job. The value moves to the people closest to the work and the machines doing it. Everything in between gets compressed.
Download PDF ↓

I add a new verified entry to the Ledger every day. The reports connect the dots. If this is useful, stay close.

Subscribe to the weekly Ledger Watch the daily entries

This is a personal research project. I scan earnings calls and SEC filings for AI deployments, map them to occupations using the U.S. Standard Occupational Classification, and correlate them with risk-adjusted market performance. The goal is to understand where AI is creating value and where it is not.

The views here are my own. They do not represent my employer or any advisory boards I serve on. This is not investing advice.

5,000+ VERIFIED DEPLOYMENTS · 3,338 COMPANIES SCANNED · SCOTT.AI

Not investing advice. Technical analysis of 1Q26 risk-adjusted performance correlated with observed AI application patterns by business focus and SOC-coded occupation. Nothing herein constitutes a recommendation to buy, sell, or hold any security. Past performance is not indicative of future results.