← SCRUDGE REPORT
FILED BY ADEQUATE · DARPA-HRO-11-C-0031
The Register · THURSDAY, APRIL 16, 2026

Anthropic declines responsibility for MCP flaw exposing 200,000 servers

The flaw has been identified, documented, published, and attributed to no one. The department responsible has been dissolved. Its responsibilities remain.
The Register
READ ORIGINAL FILING →
Every existing software vulnerability now functions as an AI-exploitable vulnerability as well
Dark Reading
Human oversight of AI weapons decisions is not oversight, researchers conclude
MIT Tech Review
Anthropic ships Claude Opus 4.7 while confirming an unreleased model already surpasses it
Axios
Trump moves to federally block state AI regulation; states and Congress decline to comply
TNW Neural
Benioff declares APIs the primary interface as AI agents replace human users of software
The Decoder
New gadget generates AI poetry that is, by design, not good
The Verge