How I built a (totally real) time machine

and how you can too.

Hi, Earth to this blog. Check, check.

I’m so glad I decided to to start a blog instead of getting a plant. Unlike a plant, a blog can go unwatered for months, only to rise back from the dead. I have, unfortunately, not had the same experience with succulents. Who knew?

But this is not a blog about my lack of consistency. It is a story of an adventure that took place in a fairly messy room on a sultry weekend, in a slow suburban neighbourhood. A certain alien species had just wrapped up a week with little plans for the weekend - until she came across a golden opportunity. An app called Lovable, which allows you to create apps from sentences made its platform free to use for the weekend. Given my enthusiasm for creating something cool, I took this as an opportunity.

To begin with, I was no don draper and I didn’t have ready ideas on what to build. But time was ticking (T-33 hours to be precise) and I had to think about something. This exploration, let me to an old idea that had been festering in my head for long. I was always a fan of https://app.radiooooo.com/, an app that lets you pick a country on a map, select a decade of choice and voila - surfaces the most popular songs of that decade. This was my biggest companion on days where my list of spotify music felt pretty uninspiring.

The discovery of this platform, against my interest for understanding history through pop culture sparked a curiosity in me - what if i could create a cabinet of songs, trends, inventions that moulded each decade? I know pudding.cool does it, so do others - and I felt pretty excited to churn one out myself! So I decided a cultural time machine: pick a country, pick a decade, and discover the songs, trends, and inventions that shaped that moment in time.

And thus began this journey - with a dream, a fire, and an arguably incoherent prompt. What came up as a result was not the magical lanscape I imagined.

T-32 hours to go

But as the famous phrase goes - “try, try until you succeed in making the best use of the free lovable credits which would otherwise cost you a fortune” - I began to iterate. And I reached a somewhat fun first output. Image below.

Now as you can see, this was a clearly better design than the previous version. I am pretty stoked at this point - but the challenge was the actual tooling of the user flow. Imagine this -

I want an arcade themed product where I select a country → pick a decade → find popular songs, trends and inventions of the decade. easy peasy, right? but the surfacing was tricky.

T-22 hours to go

But of course, building the actual flow was harder than it looked. At first, I tried connecting the backend - basically, a place where the facts and content would live (called Supabase). But this failed because the content wouldn’t surface. Also, supabase had a 20 row limit.

Then I tried plugging in OpenAI to automatically fill in the gaps when content was missing - but even that ran into errors, mostly because the access key wouldn’t connect properly. Eventually, I decided to just hardcode some sample content to keep the experience smooth - so people could still explore and get the vibe, even if the data wasn’t dynamic yet. It wasn’t perfect, but it helped me keep moving without getting stuck in setup limbo. And finally - after multiple back and forths, I got to dummy content surfacing that I felt happy about. I will soon share what the final output looked like.

T-10 hours to go

But once I got the dummy data in, I faced yet another challenge. while the design was nearly there, the UI/UX was well, bad. And no matter how I would explain changes to Lovable to fix alignment, positioning, padding - it would ignore me as I once ignored those succulents. I could feel karma hit back. It was also sunday evening and I was thin on patience. But being curt with an llm model doesn’t work, and I wanted to salvage the rest of this flow, and my sunday evening fast. So I tried and I tried refining it and went from left alignment to the somewhat ‘right’ alignment - in just 3 hours.

So there I am, exhausted, happy and super stocked to show you the final output. Are you ready? *insert drumrolls*

T-8 hours - Introducing the Culture Time Machine

Check out the live version here. The video is attached below.

T-7 hours

I just finished writing this blog. Basking in the success of having created a (cool!) website, I now retreat to the comforts of takeaway food and Superstore. What did I learn this weekend?

in an era of vibe coding, everything you dream is possible to create until you run out of credits. how do you make learning democratic, and ensure that while lovable learns to get better at listening, it doesn’t break my bank doing it?

while this was an exciting practice in creation, i do feel myself wanting to dig deeper in the weeds. what does it mean to not have to understand errors in the code? am i better for it as a creator, or worse for it as a problem solver?

Leaving this blog with some food for thought. But before that, I’d love for you to tell me what you would build out of your imagination. And if you’d like to vibe code together - let me know!

Until next time.