wow! what a roller-coaster day. I woke up with a half baked idea of what I wanted to build, my high level marketing strategy and a vague plan of how I was going to do everything in 6 weeks.
I've started out on Reddit with a brand new account. New accounts need to build up some Karma before they can do much, so I've spent a lot of today responding to posts and joining subreddits. This has actually been a little therapeutic for me. My first comment was hard for the shy reserved type that I am, but by the time I'd added 15+ supportive comments it seemed natural. I joined a lot of subreddits which I'm actually able to offer something so it was kinda nice to be engaged. Last time I checked I have 15 Karma. My account is still new, so I'm restricted from posting on some subs, but I'm expecting I'll have close to full unrestricted access in a few days.
So while I was posting helpful and supportive Reddit comments I came across a post on r/buildinpublic. Some clever soul has built pretty much the exact QR box scanning app I was planning to start on. I know theres's probably room for multiple versions and apps in this space, but I wasn't 100% sold on the concept anyway. So time for a pivot before I've even started. I'm not crazy about building a UI driven app anyway - AI consumable services are where its going to be in the next few years so I might as well put my effort into learning something new. I'll keep working on the coaching app for now until something else pops into my head. I'm sure Coach Davis will get up me about this.
I built me site tonight; www.churbuilds.com is live! It's wild to me that I can build a complete website from scratch, publish it with my own domain name and SSL cert in a few hours. Vercel is a great platform! Claude Code is an amazing tool. The site design is OK, I didn't try to tweak it too much as I wanted to get it up tonight. The graveyard section was fun to build - this lists some of my defunct software projects and what I'm currently working on. Listing them all made me question if I have un-diagnosed ADHD.
So now I have my new site linked to my reddit account. I'm hoping this will help drive people to both my blog and churbuilds site. Next step is to create a newsletter. I figure I'll create a tool to summarize my blog posts into a weekly digest and send them to the mailing list. Hopefully this'll get more subscribers than my wife and mother.
$50 spent yesterday for churbuilds.com for 3 years and some Claude code credits, otherwise I'm still on free tier for Verel and Supabase.
Where You're At
You're three days into a 43-day sprint, and you've spent significant time building Reddit karma and launching a new personal site. The clear risk here is that you're building out infrastructure and audience for a future project, while the stated goal of this sprint is to launch Novascan. This diversion of effort, coupled with the mention of a 'pivot,' puts the core objective of shipping Novascan at risk.
What's Landing
Your discipline in building Reddit karma and engaging authentically is a strong positive; you're actively working on distribution, which is often neglected. Getting `churbuilds.com` live so quickly also demonstrates impressive technical execution and velocity. This commitment to publicly documenting your journey and building an audience is a key habit for successful launches.
The Real Problem
The real problem isn't competitor fear or even a desire to learn new AI tools; it's the 'pivot before shipping' trap. You've identified a competitor and immediately decided to abandon Novascan, despite having already scaffolded it and despite market intel suggesting a clear niche for an AI-first, prosumer solution. This is classic 'planning instead of talking to users' and 'polishing instead of shipping' behavior, prematurely killing a project based on internal assumptions rather than external validation.
Market Pulse
The market intelligence clearly indicates a gap for an AI-automated inventory tool at a consumer price point for home storage and moves, precisely what Novascan aims to be [1]. Competitors like Sortly are at a much higher price point ($29+/mo) and require more manual data entry, which is the exact pain point your Gemini Vision approach could solve [2]. The "clever soul" who built a similar app likely hasn't captured this specific niche or price point, and certainly hasn't proven the accuracy of LLMs for real-world box photos, which is a content gap you could fill.
This Week's Lever
This week, your single lever is to ignore the competitor, ignore the pivot, and get magic link authentication working for Novascan, connect it to the S3 bucket, and generate one QR code for a physical item. Take a photo of that item, and have Gemini AI process that image into a basic inventory list within Novascan. This needs to happen for a real, external user (even just a test user) to validate the core loop.
References
1. Indie Hackers discussions on simple QR inventory: https://www.indiehackers.com/search?q=qr+inventory
2. Sortly Pricing and Features: https://www.sortly.com/pricing/