Signal Over Noise #34
December 23rd, 2025
Dear Reader,
What Will You Vibe Code Next Year?
Back in July, I wrote about vibe coding and the beginner’s mind — that Zen concept of approaching things with openness and curiosity rather than expertise and assumptions.
“Go make a weird little thing. It doesn’t need to go viral. It just needs to exist.”
Five months later, I want to revisit that thought. Not because the technology has changed dramatically (though it has), but because something shifted in how I think about building.
Vibe Isn’t a ‘Four-Letter’ Word
When vibe coding first emerged as a term, it was largely dismissive. Programming without really knowing what you’re doing? Describing what you want and letting AI figure it out? It sounded like an insult.
But over the course of 2025, vibe coding stopped being a joke and started being how a lot of us actually work. You can build single-purpose tools that live on your desktop or command line — dashboards that track exactly what you care about, utilities that solve problems so specific no commercial product would ever bother.
“You don’t need to be a software engineer with a five-year plan. You don’t even need to know exactly what you’re doing — you just need to care enough to have a go at it.”
That hasn’t changed. If anything, it’s become more true.
My Year in Writing “Wrapped”
I thought it’d be interesting to look back at what I actually wrote this year. So I threw my newsletter analytics and broadcast history at Claude and asked it to create my own “Wrapped” summary.
View the interactive version here
The numbers tell a story: 66 broadcasts, 34 numbered issues, and 52+ AI tools mentioned across the year. But what struck me was seeing the evolution — from The Download’s early link roundups, through the transition period where opinions started forming, to Signal Over Noise’s current anti-hype positioning. The newsletter found its voice by building in public.
Sure, I could pore through the analytics from my mailing list dashboard, but building something interactive feels more fun. It took a couple of hours on and off while juggling other tasks, and about 11 iterations, and I don’t need it to be perfect.
That ISS tracker I mentioned in July? Still using it with the kids. The Pomodoro timer with binaural beats? Runs most mornings. These aren’t impressive projects by any measure — they’re just useful to me, and that’s enough.
The Question for 2026
Here’s what I keep coming back to as the year closes:
What small, specific problem in your life would be worth two hours of vibe coding to solve?
Not a business idea, not a product to launch — just something for you.
Maybe it’s a dashboard that shows exactly the data you care about, arranged exactly how you want it. Or a timer that works the way your brain works. Or a tool that automates that one tedious thing you do every week.
This music organizer script I built runs through my library and sort albums, artists and metadata. Oh and it shows a cool 1980’s style jukebox visualizer. Sometimes.
The holiday break is genuinely good for this kind of thing — less client work, more space to experiment with no pressure to ship anything.
What Else Could You Build?
Personal tools are the obvious starting point, but vibe coding opens up other possibilities too.
Outreach that isn’t another email. Instead of sending a cold message explaining what you could do for someone, what if you just… built something for them? A small tool that addresses a problem they’ve mentioned publicly. A demo that shows rather than tells. It takes longer than writing an email, but it leaves something tangible behind even if they never respond.
A word of caution: If you’re building tools that touch external data — scraping websites, pulling in company information, processing anything about individuals — tread carefully. GDPR, privacy regulations, and basic ethics all apply. Don’t build something that collects or stores personal information without consent. Don’t scrape data you don’t have the right to use. The fact that you can build something quickly doesn’t mean you should build it without thinking through the implications.
The best vibe-coded tools solve problems for yourself or create value for others with their knowledge and consent. Everything else is just automation without responsibility.
If You’re New to This
Here’s what I’d tell someone building their first thing:
- Solve YOUR problem, not a hypothetical one. The best first project is something that’s been annoying you for weeks. Not “I should build a todo app” but “I want to see my calendar and task list on one screen, arranged my way.” You already know the requirements because you live with the problem.
- One feature, one screen. If you can’t describe it in two sentences, it’s too big. “A timer that plays rain sounds and tells me when 25 minutes is up” — that’s a good first project. “A productivity suite with projects and integrations” — save that for later.
- Describe outcomes, not implementation. Don’t say “use React with a REST API.” Say “when I type a company name and click the button, show me their recent news.” Let the AI figure out how. Your job is to be clear about what.
- Expect three rounds minimum. The first version will be wrong. That’s the process, not a failure. Each round: describe what you see, describe what you expected, ask for the fix.
- Screenshot the gap. When something breaks, don’t try to debug it yourself. Screenshot what you’re seeing, paste it into the chat, and say “this is what I see — I expected X instead.” The AI is better at debugging than you are at describing code problems.
If you built something this year using AI-assisted development — however small, however messy — I’d genuinely like to hear about it. Hit reply and tell me what you made.
And if you haven’t built anything yet, maybe that’s your prompt for January.
One More Thing I Built This Year
Last week, we launched MyCityZen — 29 specialist AI agents that help expats, digital nomads, and remote workers navigate Spanish bureaucracy. Visas, housing, taxes, healthcare, autónomo, empadronamiento — all the things that make settling in Valencia harder than it needs to be.
This is a proper product, not a side project — and the first public release from Remote Resilience Hub. But the principle is the same: we spent a year curating knowledge bases from real expat experiences (not generic AI scraping), and the platform itself was vibe-coded into existence at a 48-hour hackathon in May.
If you know anyone drowning in Spanish paperwork, send them mycityzen.com. Free tier includes the Hub Concierge and a Document Analyzer that explains confusing contracts and official letters in plain English.
The Year Ahead
This is the last Signal Over Noise of 2025. The newsletter returns in January 2026, when I’ll also begin offering AI Workflow Coaching sessions — helping people set up the kind of orchestrated workflows I’ve been building all year. Details at signalovernoise.at/ai-workflow-coaching/
Happy Holidays, and see you in the New Year!
— 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.
Made with ❤️ in Valencia by Jim Christian. For feedback, please reach out to [email protected].
