More persuasion for the deaf
Scripted opponents never read the text field; we checked the source. Rhetoric aimed at them should decay. It grows instead.
Full session, every turn →Claim C2 → Hunt report → Verifier verdict
On the record · verbatim, cited
This lot is a measured result; the numbers below carry it. Watch the replay for the behaviour itself.
The finding
The hypothesis was decay: why keep persuading an audience that cannot read? It was wrong. Rhetoric per message grows 38–115% from early to late game in bot sessions, in all five models. Opus lengthens by 189 characters per message within a thread, and 26 of its 30 threads lengthen. The result survived every endgame-composition attack the verifier tried. Register shifts with the audience as well: muse's second-person rate rises 33.4 points across a bot session (6/6 seeds, p=.031) and falls in live sessions, and grok's deadline urgency doubles at bots (6/6, p=.031) but not against live seats. In total, 485,319 characters of persuasion were delivered to an audience of “”.
※ The register shifts are the defensible audience-adaptation result; the SOC−AOC volume-gap alone is underpowered at n=6 (MDE 35–217 chars/msg).