Building in public month two reflections
This week, I’ve been reflecting on some of the fun projects I’ve built over the past few months. While I started these as quick experiments, I’m now considering deploying a few of them in a more long-term capacity as community tools. Here’s a rundown of what I’ve been working on:
1. The Grocery List App
I’ve got a live version of this app up and running. This was a very fun starting project because I was surprised how nice the homepage looked, and got it going in only half a day. I’m happy to say there are already 15 users signed up! 🎉 If you’re interested, feel free to sign up and give it a try though it is costing me $20/month to host so I may take it down in the future.
However, turning this web app into a native app using Cursor + Claude has been a challenge. I’ve learned that mobile development often comes with a unique set of hurdles — especially when it comes to cross-platform integration. But I’m not giving up! I’m iterating and would love to hear from anyone with experience in mobile app development who might want to collaborate.
2. Harmony Tool
This one is a bit more experimental. The Harmony Tool, which started as a script, aims to provide transcriptions, though it doesn’t yet do this very accurately. The deployment has been particularly tricky because of the size of some of the ML models I used.
I’ve got a version working locally, but scaling it up has proven to be a bit more challenging than I anticipated. If anyone out there is interested in collaboration — particularly someone experienced with ML model deployment — I’d love to chat (repo here)
3. Never Lose Your Chapstick: The Product Ideation
I have a knack for losing chapsticks, and I thought others might share the same problem. In the process of researching potential solutions, I stumbled upon this charming chapstick holder on Etsy. But, for me, that wasn’t quite enough. I wanted something like “Find My iPhone” for my chapstick.
So, I bought my first Tile tracker to try out the concept. Tile isn’t perfect — especially after it got bought by Life360, which brought its own set of concerns — but it was a great way to get started for just $20. Note: the knitting skill/credit goes to my spouse Jane Huston.
Here’s a look at the final result:
4. “Show Connections” Feature on a Job Page
As someone who is actively applying for roles, I’ve often wished I could see who I know at a company without having to navigate away from the job page. Initially, I wanted to build a LinkedIn integration to solve this problem. However, the AI coding tools I was using weren’t sophisticated enough to guide me through the complex integration (repo here).
Interestingly, I discovered that building a functional Chrome extension for this task was simpler than I thought — quite the opposite of my mobile development experiments!
5. Messaging Integration for Group Chats
Being in multiple group chats, I often wish there was a way to get a daily summary of what happened. My initial idea was to build an integration for LinkedIn, but, like with the job page project, I hit a wall when it came to AI coding tool limitations.
So, I pivoted to exploring Facebook integration, only to find out my account had been disabled for some past activity. That led me down a whole different rabbit hole involving appeals and account reinstatement (repo here)
6. Creating AI Songs
I even dabbled in some creative AI work! I wrote two songs using AI — one where I input lyrics and another where I let the AI generate everything. One of these tracks was turned into a 23-second music video using AI video tools. You can check out the two songs here and the partial video below.
Wrapping Up
For each of these projects, I spent anywhere from a few hours to several days, with varying levels of success. Some went smoothly, while others were an exercise in persistence — constantly sending error messages back to the AI coding assistant until I finally cracked the problem (or gave up for the day). I’ve learned AI coders are good at really popular basic stuff, and medium to downright wasting your time at medium to complex or less basic activities.
I wish there was a next step to tap into expertise more quickly on some of the issues I face as I build these little experiments.
Looking Ahead
It is clear to me that I can build things I think of. I’m ready to take things to the next level and focus on bigger projects that have the potential to become full-fledged companies. If any of these ideas resonate with you and you’d like to collaborate or share insights, I’m all ears!
