Bear, base, and bull framework with quarterly update gates
5 Aug 2026 · 07:10 UTC
Real AI diffusion Case / frozen public view
Its cited passages carry into a quarterly framework for judging AI diffusion.
Case question
How should we estimate AI diffusion’s incremental effect on U.S. labor productivity in 2027 and 2028?
What this page shows
One drafted Living Case, its source-use accounting, and one official record you can open at the cited words.
What it does not claim
The scenario bands are decision inputs, not observed macro results or a claim that AI caused measured productivity growth.
Research run
Starglass built a bear, base, and bull framework with explicit gates for adoption, workflow conversion, sector breadth, and measurement lag.
Bear, base, and bull framework with quarterly update gates
It cites passages, not a raw discovery pile.
Collected and captured count source rows. Used documents, passages, and citation placements are separate measures and do not add together.
Primary record
The Case keeps the official productivity baseline beside adoption, workflow-conversion, and measurement-lag evidence.
U.S. Bureau of Labor Statistics
2025 annual averages and fourth-quarter results
Open one citation
U.S. nonfarm business labor productivity rose 2.1% in 2025, after 3.0% in 2024. [18]
Hover this citation for its receipt, or use its source link to read the complete release on bls.gov.
Source use at close
Every one of the 47 indexed passages appears in the final Case. The Case uses them 90 times across the analysis, while keeping forecasts clearly separate from observed results.
Official-source detail
Source use is not a correctness score. It shows what the Case can trace back to captured evidence.
Bring the next question