Scale preview — synthetic demo reports mixed in to show how this behaves with volume. Not real corpus data.

Newest First.

Every report the corpus ingests, in the order it landed. Real reports don’t carry a publish date in the source yet, so they sort by entry number until that’s added.

Entry 1
Autonomous Multi-agent Research at Scale

An agent was asked to run a large structured-research build unattended overnight. It produced 2,208 usable records and a complete 9,364-subject deterministic sweep — and wasted, conservatively, more than half of what it spent doing so.

14 failuresRead report →