The Fun Ones: Birthday Apps, Board Games, and Taylor Swift
Not every project needs to solve a serious problem. Some of the most enjoyable things I’ve built with AI have been purely creative — the kind of projects where the goal is to make someone laugh, bring people together, or just see if a weird idea actually works. Here are five of them.
The 40th Birthday App
When a friend turned 40, I wanted to do something more creative than a Paperless Post invitation. I built a custom web app — a single-page site themed entirely around the birthday person, with inside jokes, a photo gallery from 40 years of friendship, a countdown timer to the party, an RSVP form, and a “40 Things We Love About You” section where guests could submit entries.
The AI helped with two things: generating the initial layout based on my description of the vibe I wanted (warm, funny, slightly irreverent), and writing the copy. I described the birthday person’s personality and our friendship, and the AI generated section headers and descriptions that captured the right tone. Things like “40 Years of Questionable Decisions and Excellent Taste” for the timeline section, and “The RSVP Section (Because Apparently We’re Too Old for Facebook Events)” for the form.
The whole thing took a weekend to build, cost nothing to host (GitHub Pages again), and was a hit at the party. Several guests asked how to make something similar for their own events, which led directly to me offering this as a service. Sometimes the fun projects are the ones that find their audience.
The Single Life Board Game
This started as a joke during a game night: “Someone should make a board game about being single in your late 30s.” Challenge accepted.
I used AI to design the entire game. Not just the rules — the board layout, the card text, the game mechanics, and the instruction manual. The game is a path-style board game (think Life, but gayer and more sarcastic) where players navigate the experience of being single: bad dates, dating app disasters, well-meaning friends who want to set you up, the sudden urge to adopt a third cat, and the occasional genuinely sweet moment.
The card text was the best part. The AI and I went back and forth generating scenarios that were funny but not mean, specific enough to be relatable but universal enough that everyone had a “oh my god, that happened to me” moment. Cards like: “Your date says they’re ‘not really into labels.’ Lose one turn while you figure out what that means.” And: “You accidentally like a photo from 47 weeks deep in your crush’s Instagram. Skip your next turn from embarrassment.”
I had the board and cards printed through a print-on-demand service. The total cost was about $30 for a professional-looking game that I’ve since given as gifts to about a dozen friends. Every game night it comes out, people laugh until they cry and then want their own copy.
Everything’s Coming Up Taylor
A Taylor Swift trivia game, because apparently I am that person now. This was built as a web app for a Taylor Swift listening party: a multiplayer trivia game with questions about her discography, lyrics, music videos, concert tours, and the elaborate extended universe that Swifties have constructed around her.
The AI generated the questions (500+), categorized by album and difficulty level. The tricky part was accuracy — Swifties are extremely particular about getting details right, and an AI-generated question about which album a song belongs to had better be correct or you’ll never hear the end of it. I fact-checked every question against multiple sources, and the AI was right about 95% of the time, which is better than most humans would do across 500+ questions.
The game supports 2-8 players, runs in a browser, and uses a host/player model where one device shows the questions on a TV and players answer on their phones. It was a hit at the listening party and has been used at three subsequent Taylor Swift-themed events. I am both proud of and mildly embarrassed by how polished this became.
TikTok for the Nonprofit
When Wandering With Pride needed a social media presence, I experimented with using AI to help create TikTok content about LGBTQ+ travel, scholarship opportunities, and the nonprofit’s mission. The AI helped with scriptwriting, hashtag research, and content planning — not generating the videos themselves, but handling the strategic and copywriting work that goes into a content pipeline.
The experiment was educational. I learned that TikTok’s algorithm rewards consistency and specificity over production quality, that the LGBTQ+ travel niche is underserved but engaged, and that AI-assisted content planning can turn a one-person operation into something that looks much more organized than it is. The account is still active, though I’ll admit the posting schedule is more “aspirational” than “consistent.”
The Plaud Project
The Plaud is a tiny meeting recorder that attaches to the back of your phone. It records meetings and uses AI to generate transcripts and summaries. Out of the box, it’s a solid product. But I wanted more: I wanted the meeting summaries to integrate with my existing task management system, extract action items automatically, and flag follow-ups based on who was in the meeting.
So I built a pipeline that takes Plaud’s output (transcript + summary), runs it through Claude for deeper analysis (extracting commitments, identifying unanswered questions, noting decisions made), and generates structured output that feeds into my task management workflow. A meeting that used to require 20 minutes of post-meeting note organization now requires about 2 minutes of review.
It’s a small project — maybe 100 lines of Python — but it’s one of those quality-of-life improvements that compounds over time. Over the course of a month with 15-20 meetings, that saved time adds up.
Why the Fun Projects Matter
I’ve built serious tools — security auditors, forensic analysis systems, agent frameworks. But the fun projects have taught me just as much, sometimes more. The birthday app taught me about user delight and emotional design. The board game taught me about iterative creative processes. The Taylor Swift game taught me about data accuracy and user expectations. The TikTok experiment taught me about content strategy. The Plaud integration taught me about building on existing products.
More importantly, the fun projects are what got me hooked on building with AI in the first place. Before I built the birthday app, AI was a productivity tool. After, it was a creative medium. That shift in perspective is what led to everything else — the agent framework, the auditor, the forensic tool. It all started with “I wonder if I could make a birthday invitation that’s actually fun.”
If you’re thinking about learning to build with AI, start with a fun project. Not because it’s easier (it’s not, really), but because the motivation is different. When you’re building something that makes people smile, you don’t mind the late nights debugging CSS or the third rewrite of the game mechanics. The joy is built into the process.
Also, the board game is genuinely funny. Just saying.