Signal Over Noise #31
December 4th, 2025
Dear Reader,
Claude Opus 4.5 dropped last week, and I haven’t stopped building since.
I’ve connected half a dozen MCP servers to my workflow, written migration scripts that moved years of journal entries into my Obsidian vault, built automation workflows that trigger other automation workflows, and mapped out a complete iOS app roadmap with ten phases before I’ve written a single line of Swift.
The new model is impressive — complex code ships in a single session, documentation practically writes itself, and problems that would have taken a weekend now take an afternoon.
Here’s what I’ve noticed, though: I’m using the time AI saves me to… use more AI.
That’s not productivity. That’s a treadmill.
The Dopamine Loop
There’s something deeply satisfying about watching capable AI turn your ideas into working systems. The friction is gone. You think it, you describe it, it exists.
Each completed project triggers a little hit of accomplishment. Ship something, feel good, start the next thing. The loop is tight, the feedback immediate, and the reward reliably hits every time.
And unlike most addictive loops, this one produces actual output — real code, working systems, documented processes. It’s not doomscrolling. It’s building.
That’s what makes it tricky to spot.
Why Systems Feel Like Progress
Building systems is safe. Systems don’t reject you. Systems don’t ask hard questions. Systems provide the dopamine hit of completion without the risk of external judgment.
And here’s the trap: systems work often looks like exactly what you should be doing.
“I’m building my personal knowledge management system” sounds responsible. “I’m creating documentation for my workflows” sounds professional. “I’m connecting my tools into an integrated stack” sounds strategic.
All of that can be true. It can also be endless refinement dressed up as productivity.
The diagnostic question: Does this project have a stopping point that isn’t “when I decide to stop”?
If the answer is no, you might be on the treadmill.
The Time-Savings Paradox
The promise of AI productivity tools is simple: save time, do more of what matters.
But “what matters” is the hard part. Without clear boundaries, saved time just becomes more building time. The efficiency gains get reinvested immediately into the next system, the next integration, the next “wouldn’t it be cool if…”
I ran the numbers on my own week:
Time AI saved me: Maybe 15-20 hours across various projects.
What I did with that time: Built more projects.
What I probably should have done with some of it: The work that actually requires other humans — client conversations, publishing, outreach.
The tools I built are legitimately useful (you can check most of them out on my GitHub here) I’m not saying any of it was wasted. But at some point, building tools to be more productive becomes a substitute for the productivity itself.
The Capability-Integration Gap
Here’s the pattern I keep seeing — in my own work and in the teams I consult with:
Capability compounds quickly with AI. You can build more, document more, automate more. The potential grows exponentially.
Integration stays constant. The rate at which you can actually deploy capability into your life or business doesn’t change just because you can build faster.
So you end up with a growing inventory of capability sitting idle — unused automations, systems for hypothetical future needs, beautiful documentation for processes you run twice a year.
This is the “integration over capability” problem I keep talking about — and it gets worse, not better, when your capability-building accelerates.
The Spectrum of Productive Activity
Not all building is equal. Here’s how I’m thinking about it:
Actually Moving Things Forward:
- Work with an external deadline or accountability
- Projects where “done” means someone else responded, bought, or used it
- Building something a specific person asked for
Useful but Potentially Endless:
- Internal systems that improve your workflow
- Automation for recurring tasks
- Documentation and organization
The Treadmill:
- Building for hypothetical future needs
- Systems that support other systems
- “Getting ready” that never transitions to “doing”
- Optimization of things that already work
The middle category is where it gets tricky. That work is genuinely valuable — right up until it becomes the only work you do.
What I’m Trying Instead
I don’t have this figured out. But here’s what I’m experimenting with:
1. The External Touchpoint Rule
Before a project counts as “done,” it needs one external touchpoint. Someone who isn’t me has to see it, respond to it, or use it.
This doesn’t mean everything needs to be public. It means completion requires moving outside my own head.
2. Time-Boxing the Build
AI makes it easy to keep going. “One more feature” costs almost nothing when the feature takes ten minutes.
But the total time still adds up. I’m experimenting with hard stops: two hours on internal tools, then move to something external-facing. The tool doesn’t have to be perfect. It has to be done enough.
3. The “So What?” Check
Before starting something new: What does this enable that I’m currently blocked on?
If the answer is vague — “it would be nice to have” or “I might need this eventually” — it goes on a list instead of becoming today’s project.
The Actual Point
AI has made building cheap. Ideas become reality faster than ever.
That’s genuinely great. I’ve shipped more in the last month — heck, the last ten days even — than in some entire quarters.
But cheaper building doesn’t automatically mean better outcomes. It might just mean more building. And more building, by itself, isn’t the goal.
The goal is whatever you were trying to accomplish before you started building.
If you’re like me, it’s worth occasionally stepping back and asking: Am I still moving toward that? Or did “being productive” become the destination?
Your Turn
What have you built recently that’s genuinely useful versus what felt productive in the moment?
I’m not asking to judge — I’m asking because I think a lot of us are navigating this same thing right now. The tools got dramatically better, and nobody handed us a manual for “how to not just build constantly.”
Hit reply if you’ve got thoughts. I’m genuinely curious how others are handling it.
Until next time,
Jim
€9.00
The AI Writing Field Guide
Stop Sounding Like Everyone Else Using AI
The 5-step system for teaching AI what your voice actually sounds like — not… Read more
Signal Over Noise is weekly, reader-first publication on AI “without the hype” published by Jim Christian. If you’ve been forwarded this issue, you can subscribe for free: go.signalovernoise.at. You can also Join the free Signal Over Noise Community on Skool where I go into much more detail.
Made with ❤️ in Valencia by Jim Christian. For feedback, please reach out to [email protected].
