I Do Run AI Agents Overnight. Here's What Actually Matters.
Boundaries, supervision, evidence, and why unattended execution is not the same thing as unbounded autonomy.
Petie Clark — Applied AI Systems
Applied AI systems that survive contact with the business.
Field notes on agent orchestration, evaluation, authority boundaries, enterprise integration, deployment, and production operations.
Get new posts in your inbox
No spam. No marketing. Just the writing.
/start-here
How I think about governed systems
Why I Prefer One Clear Source of Truth
Why tools can multiply while authority, state, and accountability still need one dependable home.
Why I Write Down the Rollback Before the Change
A deployment is not ready merely because the forward path looks convincing.
The Case for a Smaller Blast Radius
Production systems become trustworthy through bounded change, not theatrical confidence.
A Service Is Not Done When It Starts
Getting a service to start is a satisfying milestone. The image pulls, the process binds, and the first request completes. It’s also the moment when a temporary experiment can quietly become a permanent obligation.
4 min read
Read →
The Maintenance Window Is Part of the Design
A maintenance window isn’t merely a polite time to make a change. It is a finite engineering resource. The design has to fit inside it with room for observation, correction, and retreat. When the change plan consumes the whole window on paper, the failure has already started; the clock just hasn’...
4 min read
Read →
What I Learned From Rebuilding Twice
Rebuilding a system sounds like repeating the build. It rarely is. The first version accumulated decisions in the order I encountered them. A rebuild asks me to reproduce the result after that order, and much of the reasoning behind it, has disappeared.
4 min read
Read →
A Better Way to End a Work Session
A work session usually ends before the work does. The clock wins, attention gets thin, or the next obligation arrives. That’s normal. What causes trouble is leaving an unfinished system in a state that only makes sense to the person who has been staring at it for three hours.
4 min read
Read →
Why I Prefer One Clear Source of Truth
An agent can receive instructions from almost anywhere. A chat message describes the task. A scheduler supplies parameters. A repository contains conventions. A wrapper adds defaults. A saved prompt contributes another set of rules. Each source is convenient in isolation.
4 min read
Read →
When Convenience Creates a Second System
A container management screen can make one small change feel harmless. Select a running service, adjust an environment value, click redeploy, and the new behavior appears. The repository remains untouched. No syntax to remember, no file to open, no review needed.
4 min read
Read →
The Most Useful Automation Is Easy to Ignore
One of the local automations I value most doesn’t have a dashboard. It watches a folder, waits for a file to finish arriving, applies a predictable transformation, and places the result where the next ordinary tool expects it.
3 min read
Read →
I Do Not Need Every Tool to Be Smart
The most useful place for intelligence in a data tool is usually a small box labeled “these two records might be the same.” Everything around that box can be ordinary comparison code.
4 min read
Read →
The First Five Minutes of an Incident
The first five minutes of an incident are rarely long enough to diagnose it. They are long enough to damage the evidence, widen the uncertainty, and create three competing versions of what happened.
4 min read
Read →
A Quiet System Is Not Always Healthy
Low-traffic systems age in place. They can sit for weeks looking perfectly composed while credentials expire, permissions drift, indexes lose contact with their sources, and background workers retain assumptions nobody has exercised since the last upgrade.
4 min read
Read →