The AI Era: Time to Play

I was on my train commute, standing at my favorite spot near the door, staring at my phone.

Debugging a web app I had been building over the last few weekends.

My workflow is simple. Identify an interesting problem, go back and forth with multiple LLMs, usually ChatGPT or Gemini, clarifying my thoughts around how to resolve it with a product.

And when I am ready to start building, I use Replit agent to build and deploy it.

This product was almost ready but wasn't working. And Replit agent (actually, it was me!) couldn't resolve the issue. I took it to Gemini, which listed a few probable causes. On the train, I was going back and forth between Gemini, Replit, and Google Cloud Console, implementing and testing the debugs.

A little later, as I exited at my stop, I started to deploy the web app into production mode. Still on my phone.

And that's when the hilariousness of the situation hit me. I was laughing as I made my way out.

Here's the thing.

I have no background in coding. Never done it. Most of the terms I used earlier, I had no clue what they meant.

And even now, my understanding is not that of a technical person.

It's more of a jack of all, I don't have a hammer, so I will use a kitchen spoon to whack the nail in, kind of perspective.

I frequently get the question from folks who are in the same boat I was in.

How do I get started with AI?

What course should I take? What should I learn?

I think that they are asking the wrong follow-up questions. My short answer is, "Don't worry about learning." Let your passion or expertise lead the way.

Start doing something with AI in an area that excites you or one that you are good at. The learning happens automatically. The journey will be fun.

Building products is not my forte. But I am good at identifying problems and figuring out a solution. And I have huge patience for trying a ton of things to figure it out.

That's what I bring to the table to my AI partner.

Begin with something you are passionate about or skilled at.

You have unique strengths, whether a subject, skill, mindset, or just something that excites you. That's what you start working on along with AI. With no end goal other than playing with it and having fun.

Back to my product.

I like to build products which address a problem I am facing. Right now, I am building a web app to search my saved YouTube playlists and videos.

YouTube is my second brain for video content.

I have a couple of thousand videos saved in my YouTube playlists. They are curated over several years on topics that interest me: marketing, behavior design, habits, finance, movie clips, music, and standup comedy.

They are of high value to me because a past version of me came across them and decided they would be worth saving. Now YouTube doesn't let you restrict your search to your saved videos. The search happens by default across YouTube.

But I am only interested in my curated videos. It's a pain to find them as I need to sift through many videos that YouTube surfaces.

So all my product will do is help me search only within my saved playlists and surface those links for me to view back in YouTube.

Yes, it's super-niche. You probably don't need it. I never set out to build the next Google. I am happy playing in my backyard with the latest shiny toy.

AI is empowering. It lets me do things I could not imagine being able to do before. And that's what excites me.

The best way to start? Just jump in.

Anywhere you start is a good spot to begin.

I wrote this back in July 2025 but never hit publish. Felt like too much ado over nothing.

Here's what changed: I'm still at it. Still building. Still playing.

But the tools have evolved—I've moved from Replit to Claude Code and spend most of my time in the terminal now. Something I would've shuddered at a few months back.

But the approach? Same.

I am not a coder. I am becoming code literate by building things I actually need.

That YouTube playlist search app?

It's live at playlistseek.com. I use it regularly for my video research.

And the learning? Happened automatically.

So if you are still wondering —don't overthink it.

Find something you want to build. Jump in.

The rest follows.