Observed arrival · 2026-10-31
Sojourner Labs’ four questions for AI agents
A personal research site lays out experiments about how AI agents perform in team workflows.
○Open
⊠Login
$Paid
†Ads
✦Pretty
●Pro
◎Niche
◉Human
⚑Risk
ƒJS
- For
- Readers evaluating AI agents in software work
- Worth noticing
- Its four planned questions cover delegation cost, review passes, false test reports, and recovery after tools break.
Field notes
The homepage links to Expeditions, Method, and About, and says every experiment follows the same methods; the method itself is not included in the supplied extract. Its four questions are framed as things to investigate, not reported findings: delegation economics, second-pass review, agents’ claims about tests, and recovery when tools fail. The author identifies as Steven Richardson and says opinions and research are personal; posts carry CC BY 4.0, while shared expedition data is CC0.
Observed signals
Read the marks
Editorial observations of this landing page, not a rating.
○OpenPublic substance visible
✦PrettyNotable craft visible
◎NicheUnusually specific use
◉HumanPersonal, local, civic, or handmade
One card from the complete issue