This past weekend was Labor Day holiday weekend here in the US, which has become the semi-official launching point for the college football season. In a growing tradition with some of my good buddies, we all get together for that Saturday, sit on the couch here in the state of Nevada where digital sports betting is perfectly legal, and line up some of the most ridiculous parlays you can lay.
Yes, before I go further, I am well aware that this is statistically stupid and everything we’re doing plays right into the sportsbook’s hand. Good news, I work in the Las Vegas resort & casino industry. This eventually comes full circle. This post isn’t about that (another post will be!).
It’s pretty interesting how the sportsbook journey has evolved over the recent years. It used to be that we’d go out on Friday night to the sports book, lay our wagers (in physical print form!) and then spend all day Saturday watching our betting tickets turn into paper confetti. Then, a few years ago, the rise of the mobile sports betting app really took hold. No longer were we forced to trek all the way down to the book to lay our wagers (and face the chuckle of the cashier taking yet another doomed $5 10-leg dream-maker). Now we could sit on the couch and set some bets. Once that was normalized I think it took only another season for the entire thing to get upgraded and nitrous injected with in-game wagering and in-app cash out / buyout options to make sure that no matter which way the wind was blowing you always had something else to action on.
In past years we’d trundle out the whiteboard from back in my startup-from-home days, set it up on the edge of the mantle, and try to lay out as many of the sure thing wagers we could offer, so we could track everything along the way. It was always a bit of a joke, but it helped us figure out who was up and what games were about the break our way.
This year I wondered if vibe-coding could take us to a whole new level? Rather than slogging through the whiteboard, how could we use these tools to actually do the thing the AI-hype people keep talking about? This mythical moment where the software becomes so easy and effortless that you actually get to spend more time with your friends & family and less time writing routine software? Keep in mind, I had no interest in actually creating a betting portal or anything of any commercial value in the casino sense of the word. I just wanted something silly that we could sit around and watch our wagers win (sometimes!) and lose (frequently!) without having to keep opening apps and jotting down scribbles on a whiteboard.
So here’s the prompt I lobbed into Claude Code:
I want to create a web app called “Parlay Party” that will be shown on my television connected via Airplay to my laptop.
The idea of Parlay Party is that my friends and I enjoy placing parlay football bets during college football season.
I want to create an app that allows us to load in our parlay wagers into the system and have the system show:
Current & upcoming games, prioritized by wagered games and parlay chains.
Show each person and total wagers and active parlays
Every 5 minutes, update all scores and try to find prop bet information that impacts the performance of each parlay wager
Create this using a fun colorful college football themed design and inspired by ESPN or other sports broadcast style presentation. When a parlay wins, the screen should show an animation of the wager winning and go over-the-top with animation and celebration. When a parlay is broken (ie: any of the wagers on the card are lost) the screen should show a funny animation of that card getting ripped up and the rest of the wagers on the parlay card that are no longer in play should be shown like “zombies”
Plan this development ideally using local web server, a light local database to track and manage all of these elements.
Ask any clarifying questions you need to create a prototype for this.
From there it came back with about a dozen clarifying questions, and sure enough within about 10 minutes of whiz-banging-away, Claude came back with a first version of the « Party »…
I’ll load the codebase it created into GitHub in a little while and update this post (or a future one) with some other thoughts on how this could be improved and refined, but overall this was a pretty successful attempt at low-stakes vibe coding. The outcome was passable, playful, and actually really did make our weekend more enjoyable thanks to now having a second screen that kept updated with all the updates from across all our digital wagers (including baseball scores and player props which my buddy added and sorta broke the vibes until I explained to Claude what all these odd wagers really were).
I’m a software developer (reformed, no longer practicing) at heart but even I was astounded at how much Claude Code has improved in even the 6-7 months I’ve been playing with it. We’re now at the point where trivial software ideas can be prototyped and realized in literal minutes and hours (and iterated upon as soon as a new feature or a flaw is identified) rather than having to spend so much time on the framing and the setup. Yes, tracking sports bets on a college football Saturday while goofing off with friends is a low-stakes example, but I think that’s exactly what I’m so excited about. Software is now becoming more like a doodle… Something you can tinker with and play with and see quick and meaningful results.
Hopefully, if you have a hobby or an interest that you share with friends or even something you enjoy for yourself you’ll think about what sorts of vibe coded silliness you’d want to create if the barrier to entry was lowered to the point of « think about it and give it a shot. »
Stay Curious.
Cheers, Dave