San Francisco
Four things I found by measuring what usually gets argued about.
15 of the 39 most-taught persuasion techniques have no peer-reviewed study behind them. 11,785 records screened across 7 databases, 44 extracted. The protocol and the deviation log were committed before the results, so the work can be checked against what I said I would do. closing-evidence-atlas
A colour that clears WCAG 4.5:1 against a dark ground lands at APCA Lc 35, under the floor for text at any size.
The standard false-passes dark interfaces. One file, no dependencies, and --selftest reproduces four published reference values before you trust a number out of it.
apca-gate
A citation pass on my own taxonomy corrected three of my errors and pulled two citations that do not resolve to any real publication. Assessment 001 carries two adverse findings and no composite score, because only one of six criteria was settleable from public evidence. cognitive-attack-surface
I pre-registered a four-arm trial on my own cold-outreach openers and published the result regardless of which way it went. 400 decisions, 12 weeks. sales-instrument
- mechanism-detector — names which psychological mechanism an AI conversation is running, turn by turn. Labels manipulation and has no generation endpoint.
- pam1-bandit — a fruit-fly mushroom-body model playing a slot machine, and a sweep for which payout schedule hooks it hardest.
- lsm-discovery — a Python replication of Pennebaker's Language Style Matching, validated on persuasive dialogue.
- clearframe-public — names the persuasion techniques in a headline and steelmans both sides without ideology labels.
Nine iOS apps on the App Store, written solo in SwiftUI. weight-and-watt computes your maintenance number from your own weight trend rather than from a formula. WeighIt is decision analysis built on Analysis of Competing Hypotheses.