Blog
Field notes
Short, practical writing on CyberArk, PowerShell automation, and AI workflows that hold up in real environments. First posts are in the works.
Recent posts
Do You Blindly Trust AI? What Watching AI Install Things Taught Me About Supply Chain Risk
AI coding agents make installing packages effortless. That convenience hides the security decisions AI is making on your behalf. A Remotion tutorial, a transcription task, and two unwanted ffmpeg installs showed me exactly how that plays out.
Constraints Beat Prompts Why the "Perfect Prompt" Myth Is Costing You
Prompt engineering is dead. What matters now is constraints and the evidence proves it cuts through the noise better than tweaking your wording ever will.
Consistency Over Cleverness: Why Your Best One-Shot Prompt Fails Next Time
One-shot prompting works — until it doesn't. Here's why structure beats prompt magic.
Is Copilot Even in the Game?
Copilot's usage numbers are as bad as people say. That's not really the point. Most people, on any AI tool, skip the one skill that actually matters: controlling the context, not polishing the prompt.
Most AI problems I see do not start with the model
Most AI drift doesn't start with the model, it starts with the workspace around it. Here's how ICM and MWP use folder structure instead of bigger prompts to keep agent work interpretable and under control.