OTDR Trace Reading: Three Mistakes Even Senior Field Techs Make
I've watched a fifteen-year fiber technician misread a ghost reflection as a real splice loss. He took the OTDR screen at face value, called the trace a bad
I've watched a fifteen-year fiber technician misread a ghost reflection as a real splice loss. He took the OTDR screen at face value, called the trace a bad
Tutorials show RAG working at 95% recall on toy datasets. Production systems hit 60%. The gap isn't the technology — it's three implementation choices that look harmless and aren't.
Open any 400ZR datasheet and you'll find a reach number. 120 km. Sometimes 80 km. Sometimes 140 km on the more aggressive specs. That number is a clean-link
Model Context Protocol gives AI systems tool access. It also creates a new attack surface: tool poisoning, indirect prompt injection, and privilege escalation through tool chaining. Almost nobody is scanning for it.
The MSA — Multi-Source Agreement — is a document. It is a good document. It defines the mechanical envelope, the pinout, the power envelope, and the management interface for a class of
I used to think stopping meant falling behind. It took me years, a crash, and a lot of uncomfortable honesty to understand that rest is not the opposite of ambition. Sometimes it is the only reason you survive it.
I closed 47 browser tabs on Tuesday. I know it was 47 because the extension that counts them announced it like a milestone. By Wednesday morning I had 51 open
License choice is a product decision, not just a legal formality. Apache 2.0 over MIT isn't paranoia — it's the patent protection clause, which matters specifically for security software.
Pluggable ZR did to metro DCI what nobody expected at this pace. Between roughly 2022 and 2025, transponder shelves got pulled out of metro builds across most of the hyperscalers,
The stack is the side effect, not the project. Two principles drive every decision — keep it local, and move every model to one I trained myself. Everything else is downstream.