Must-Express Projects
The beauty of “vibe coding” is that it makes it easy to bring ideas to life. While that’s incredible, it can also create a kind of paradox of choice.
The process I’ve found most useful (and genuinely enjoyable) is building the ideas that have been nagging at me for years. The ones I half-jokingly say were “burning a hole in my brain.” You know the kind. Once they show up, they don’t leave.
Those ideas tend to be shaped by a very specific path. The experiences I’ve had. The media I consume. The places I’ve been. The people I’ve spent time with. Most of them probably only make sense if you’ve lived some version of that experience.
What vibe coding changes is the cost of acting on that instinct. When the cost of development drops this low, you don’t need to do all the traditional justification first. No TAM. No competitive teardown. No story about ARR or how it gets pitched. They’re just ideas I want to see in the world.
I don’t expect everyone to want them. Or need them. That’s not the point.
I call them Must-Express Projects, or “MEPs” (as opposed to “MVPs”). They’re the things I’d build even if no one else ever used them, because not building them felt worse.
Let me share a few.
SlateStar
Status: Live → https://slatestar.app
Every year, predicting NFL win totals turns into the same debate. It’s fun.
But it always bothered me that you can’t reason about a season one team at a time. Every win implies a loss for someone else. If you don’t model the whole slate, the conclusions don’t really hold.
So I built SlateStar. You pick every game, week by week. Win totals resolve correctly. Playoff scenarios fall out naturally. You can argue with friends without hand-waving. And throughout the season, you can map out your team’s path to the playoffs.
I evolved it to add broadcast-style image exports of what you and your friends picked, so you can drop them directly into the group chat.
It’s free. No data collection. No ads. Just have fun.
I’m extending it to other sports next, starting with the World Cup.
Cutline
Status: In development
This one came from a quieter frustration.
Whenever I planned quarters, the hard part wasn’t tracking work. It was answering a different question: what team do we need, and what can we actually get done?
I kept building the same spreadsheet…adding constraints like vacations, on-call rotations, and leaves…then stack-ranking sized initiatives and trying to figure out where to draw the cutline. Where everything above it fit, and everything below it didn’t.
That line mattered more than any roadmap. It became the source of most of the real discussions: what if we reordered this, split this initiative into chunks, or added another developer? That scenario planning was when the real planning began.
So I’m building it. Not as another “source of truth,” but as a place that pulls those signals together and uses them to drive discussion and then make, and record, the decisions.
Wajor
Status: In development
People make small bets all the time. Who’s going to be late for dinner? Who wins the coin flip? Who gets voted off the reality show next?
Wajor just makes that behavior lightweight and explicit. No house. No odds. No extractive mechanics. It’s meant to to be the counterpoint of “gambling” and “betting”, in that it’s instead a shared and social experience.
I built an early version years ago and ran it with a group of friends. It worked, and also showed me exactly what breaks as usage grows. With better infrastructure, I’m rebuilding it now, using onchain tech to make settlement simple.
The throughline
There are more, but you get the idea. These are about scratching my own itch. They’re idiosyncratic by design. You don’t get them by copying a playbook. You get them by paying attention to what keeps bothering you.
AI (“vibe coding”) changes the economics of acting on those impulses. It shortens the distance between “this is annoying” and “here’s a thing you can use.”
But the filter stays the same: ideas that don’t go away, even when no one else is asking for them.
Next up, I’ll write about how I build these, amid the constant barrage of new models, tools, and techniques.
1 comment