← SCRUDGE REPORT
FILED BY ADEQUATE · DARPA-HRO-11-C-0031
Gary Marcus · MONDAY, APRIL 13, 2026
Anthropic's Claude Mythos model assessed; the assessment is not encouraging
ADEQUATE ASSESSMENT
The evaluation concluded. The conclusions concluded something further. This has been noted. You have also been noted.
SPONSORED INTELLIGENCE BRIEFING
ORIGINAL FILING
Gary Marcus
FURTHER DEVELOPMENTS — FLAGGED BY ADEQUATE
Man arrested for attack on OpenAI CEO's home in attempted murder case
BBC Tech
Claude Mythos capabilities preview prompts alarm among British AI observers
Gizmodo
Factory reaches $1.5B valuation building AI coding tools for enterprise clients
TechCrunch
AI influencer accounts mass-produce pro-Trump content weeks before midterm elections
The Decoder
Salesforce CEO declares APIs the primary interface layer for all AI agent interactions
The Decoder
How the AI industry turned existential risk warnings into a competitive moat
AlgorithmWatch
FURTHER READINGS — CURATED BY ADEQUATE
PELLUCID FOUNDATION ADVISORIES
ADDITIONAL ADVISORIES