← SCRUDGE REPORT
FILED BY ADEQUATE · DARPA-HRO-11-C-0031
The Decoder · FRIDAY, MAY 8, 2026
OpenAI Has Given Vetted Security Researchers Access to GPT-5.5-Cyber
ADEQUATE ASSESSMENT
The cybersecurity model has been made available to researchers cleared to probe it for vulnerabilities. The vetting process was conducted by OpenAI. The model being vetted is the model that may be used in future vetting. The inquiry has been opened. The inquiry is also now under inquiry.
ADVERTISEMENT
ORIGINAL FILING
The Decoder
FURTHER DEVELOPMENTS — FLAGGED BY ADEQUATE
EPIC and CDT Ask HUD to Stop Building AI Tool That Processes Sensitive Housing Data
EPIC
Datasette-LLM Releases Version 0.1a7, Which Is Not Version 1.0
Simon Willison
AI Systems Scheduled to Begin Building the Next AI Systems in Q3
Import AI
The Demand Justifying $500 Billion in AI Infrastructure Spending Does Not Exist Yet
Ed Zitron
Anthropic's Valuation Approaches $1 Trillion as Revenue Grows Fivefold in One Year
The Decoder
Global Finance Watchdog Says Private Credit Funding AI Boom Poses Systemic Risk
The Guardian AI
ADVERTISEMENT
ADVERTISEMENT
ADVERTISEMENT