Observed arrival · 2026-10-13
Simulated Research’s AI Field Reports
Independent researcher Rick Willison publishes dated reports on AI-agent behavior, alignment probes, and reliability tests.
○Open
⊠Login
$Paid
†Ads
✦Pretty
●Pro
◎Niche
◉Human
⚑Risk
ƒJS
- For
- AI alignment and agent-evaluation researchers
- Worth noticing
- One report explicitly treats a six-before/zero-after audit split as a lead, not a result.
Field notes
The report index links dated session write-ups, including an audit that sent fifty variations of an ordinary frustrated message to a model and classified its replies. A separate report logs seven accuracy failures alongside a probe for experience claims, while cautioning that the observed audit split came from one session. The site also describes a live wish machine whose rules bar agents from asking for or granting wishes.
Observed signals
Read the marks
Editorial observations of this landing page, not a rating.
○OpenPublic substance visible
✦PrettyNotable craft visible
●ProPolished or operationally mature
◎NicheUnusually specific use
◉HumanPersonal, local, civic, or handmade
One card from the complete issue