RC RANDOM CHAOS

Articles

Long-form writing on tech, culture, and the edges of the internet.

The Open Courts Act exposes what PACER fees hid
data governanceAI safety

The Open Courts Act exposes what PACER fees hid

PACER's per-page fee was an accidental privacy brake. Making court records free is right - but only if redaction, governed bulk access, and security replace it.

18 min read
Trust does not carry forward
systems drifthallucination

Trust does not carry forward

GPT55 hallucinated three times more than GLM52 on identical prompts. The cause is systemic: systems resolve references without revalidating their content.

6 min read
ZFS does not keep your data secure
ZFSUbiquiti NAS

ZFS does not keep your data secure

ZFS gives Ubiquiti's enterprise NAS strong data integrity, but RAID-Z, snapshots, and cloud accounts each fail in ways the filesystem won't fix.

7 min read
It measures conformity, not security
grapheneosvolkswagen

It measures conformity, not security

Volkswagen blocks GrapheneOS while admitting baseline devices. The gate measures configuration identity, not security posture.

8 min read
Social engineering weaponized an Anthropic model
social engineeringidentity and access control

Social engineering weaponized an Anthropic model

The Anthropic Mythos event involving a Korean telecom was a failure of identity and access control against known social engineering vectors, not a data leak.

7 min read
speed ships the flaw
ai securitysecure development

speed ships the flaw

AI generates code, config, and infrastructure faster than validation can check it. The gap between production rate and validation rate is the attack surface.

8 min read
The cache we enabled was a 25% surcharge.
prompt cachingclaude code cost control

The cache we enabled was a 25% surcharge.

A broken prompt-cache config charged Foundry a 25% input premium for 11 days. The fix, the batch ratio, and why RAM prices set your token bill.

7 min read
The same AI you're shipping wrote the malware
LLM deploymentAI security

The same AI you're shipping wrote the malware

10,000 trojan GitHub repos weren't a malware breakthrough - they prove LLM safety lives in the model while abuse happens in the unguarded pipeline.

11 min read
Zero-Touch OAuth strips the 2025-06-18 MCP mandate
MCPOAuth 2.1

Zero-Touch OAuth strips the 2025-06-18 MCP mandate

Zero-Touch OAuth for MCP fails as a trust-on-first-use design: unauthenticated dynamic client registration and unbound bearer tokens enable session hijack.

7 min read
GitHub's scanners cleared 10,000 trojan repos
supply chain securitygithub security

GitHub's scanners cleared 10,000 trojan repos

10,000 GitHub repositories distributed trojan malware because platform presence was treated as validation. The control was assumed, not enforced.

7 min read
MITRE already filed your detection bypass as AML.T0015
adversarial-mlml-edr

MITRE already filed your detection bypass as AML.T0015

ML malware detection is a deterministic classifier with a mappable decision boundary. Attackers exploit its learned bias. That demands more engineering.

7 min read
Certified is not secure
ai-orchestrationsecurity-architecture

Certified is not secure

Volkswagen blocking GrapheneOS shows what it costs when one attestation flag replaces a real risk decision, and why orchestration beats a longer blocklist.

11 min read