Tech Notes

I tried making a quick ad video (10-15 seconds) today using Google Gemini Pro for my productivity app: Aura Tracker: Habits & Goals, but it kept throwing me off with wrong results, which got me really frustrated and I finally gave up. After thinking about it, I realized this is a common issue with AI tools. Many AI tools make it seem like using them is super fast and easy, with influencers even claiming they can quickly create amazing results with a specific tool. So, when the tool keeps failing, frustration builds up, and you’re more likely to give up quickly than if you’d started with a manual approach. ...

AuraTrackr’s testing strategy was built around realistic behavioral simulation. Since the app’s Insights and Analytics rely on pattern recognition, like correlations between sleep, exercise, and deep work—manual testing only works if I can simulate weeks of human behavior in seconds. This isn’t about checking whether charts render. It’s about validating that insights remain meaningful across different psycholoical profiles. 1. The Seeding Engine: Beyond Random Data Most apps seed test data with placeholder entries. That doesn’t work here. ...

After 20+ years of building software for big corps, I finally shipped something that was fully mine. Building something from scratch is a journey of both obsession and philosophy. Over the last 1.5 months, I’ve poured my two decades of engineering experience into a single vision - to make Aura Tracker: Habits & Goals help you transform your life. I didn’t want to build just another habit / goal tracking app. I started with a core belief that transforming your life requires more than a streak counter. It requires transforming your identity and building resilient systems. I also wanted one time purchase, offline first, no ads, no account creation, no subscriptions. When I looked at the market, I couldn’t find anything that seamlessly combined that entire flow. So, I decided to build it myself. ...

I’ve always been curious about Google’s auto-correct feature. Let’s face it, it looks ultra cool, especially when we type in some vague, disjointed letters and Google suggests exactly what we had in mind. It almost feels like Google is reading our thoughts! With some free time on my hands, I decided to dig into how it actually works. Of course, I don’t have the monstrous server farms that Google has at its disposal to build an equivalent web-scale solution. Hence I thought of and implementing a version more suited to a local/desktop setup. ...