Loading the edition…
· Past edition · 10 stories
Retrospective edition reconstructed from later captures. Paper versions may include later updates; coverage labels describe the evidence used.
Research & academia
A benchmark tests building software from interactive references.
Why it matters: Interactive application references can become executable tasks for training and evaluation.
Research & academia
A benchmark measures compliance when enterprise assistants face pressure.
Open each story in the reader. Summaries are shown when available; some sources provide an abstract only. With an account you can keep your own notes.
Why it matters: Enterprise agents must follow rules even when user pressure conflicts with them.