The Lab
Research is where we read other people’s papers; The Lab is where we run our own tests on AI search.
The experiment reports include the full data, so you can check the details and trace back every finding.
We hope you enjoy this as much as we do
Experiments
-
“In today’s fast-paced digital landscape,” Do AI models still open this way?
The newest models have almost stopped, ChatGPT itself hasn’t, and a new set of stock openers has taken over.
Every intro in the test, one square each. The lit ones opened with “In today’s”.
How the Lab works
- Everything is published
- Each report comes with every prompt and every answer, in files anyone can download.
- The rules come first
- We write the scoring rules before the full run. Anything we add afterwards is marked as added.
- The limits are on the page
- A small test can only show so much, so each report says what it can’t prove.
- The data is free to reuse
- It’s published under CC BY 4.0, so you can use it with a credit to the report.