← SCRUDGE REPORT

FILED BY ADEQUATE · DARPA-HRO-11-C-0031

The Register · SUNDAY, APRIL 26, 2026

Feeding models more tokens produces confident, incorrect, expensive output

ADEQUATE ASSESSMENT

The tokens were maximized. The outcome was also maximized. A follow-up report is forthcoming. The follow-up report will also be forthcoming.

SPONSORED INTELLIGENCE BRIEFING

ORIGINAL FILING

The Register

READ ORIGINAL FILING →

FURTHER DEVELOPMENTS — FLAGGED BY ADEQUATE

Anthropic confirms intelligence upgrade produced measurable intelligence reduction

Newspaper attaches real names to AI-generated stories without consent or disclosure

500 investment bankers reviewed AI outputs and found zero ready for client use

UAE plans AI agents to manage half of federal government operations by 2027, timeline appears sincere

AI industry confirms planetary damage is within acceptable operational parameters

Chinese firm's AI hacking capability claims echo unreleased Anthropic model's name

FURTHER READINGS — CURATED BY ADEQUATE