RC RANDOM CHAOS

Articles

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

A model brute-forced Linear A
security-through-obscuritycryptanalysis

A model brute-forced Linear A

A Linear A decipherment claim exposes the lesson defenders miss: complexity is not security, and scaled pattern recognition defeats obscurity.

6 min read
Google IPv6 crossed 50%, your IPv4 controls didn't follow
ipv6 securitydetection engineering

Google IPv6 crossed 50%, your IPv4 controls didn't follow

Google's IPv6 traffic crossed 50 percent. IPv4-era controls don't see IPv6 - the kernel, NDP, scanning, and telemetry reality for defenders.

6 min read
Google killed io_uring fleet-wide in 2023
io_uringepoll

Google killed io_uring fleet-wide in 2023

io_uring runs file and network operations off the syscall path, blinding seccomp, auditd, and EDR, while epoll stays observable to defenders.

7 min read
Keep the hard part
AI workflow designcognitive offloading

Keep the hard part

AI doesn't erode your problem-solving skills-offloading the reasoning does. Any intelligence atrophies without use; the fix is design, not avoidance.

11 min read
Linux kernel deleted strncpy across 360 patches
linux kernelmemory safety

Linux kernel deleted strncpy across 360 patches

Linux removed strncpy across 360 patches over six years. The exposure: a bounded write primitive used as a safety control it never implemented.

8 min read
Social engineering is a misconfiguration
social engineeringidentity security

Social engineering is a misconfiguration

Human error in identity workflows is a misconfiguration, not incompetence - how Scattered Spider, 0ktapus and MFA fatigue exploit the validation gap.

7 min read
The channel trusted the sender
mobile carrier securityaccess control

The channel trusted the sender

An unauthorized alert reached phones across Brazil. The confirmed finding is one control: sender authorization at the injection point did not hold.

7 min read
The green check mark proves nothing
AI code securitysecure code review

The green check mark proves nothing

Accepting AI-generated code because it works extends your trust boundary to an unvetted source running under your identity. Function is not authorization.

9 min read
The same-origin policy is not protecting your API
corsapi security

The same-origin policy is not protecting your API

A permissive CORS header delegates the read decision to the requester, letting attacker script read authenticated responses through the victim's own browser.

7 min read
2023 mistakes an IP address for a passport
identity verificationnetwork security

2023 mistakes an IP address for a passport

Forcing real ID on all internet traffic relocates an unsolved identity problem to a layer that cannot verify the subject and creates a higher value target.

8 min read
2026's AI failures aren't model failures
LLM engineeringAI orchestration

2026's AI failures aren't model failures

AI deployments fail at orchestration, not capability. Building validated pipelines around the model - not completing the task - is the real job.

10 min read
A bypassed control is worse than no control
web application securitycontrol bypass

A bypassed control is worse than no control

CSSQuake is a demonstrable bypass of web application protections through trust manipulation and control boundary erosion. A breakdown of what failed and why.

7 min read