For weeks, "Build AI prototype" sat on my to-do list. Every single day, I'd find other, easier things to do – read the latest AI news, watch tutorials, take online courses. Learning about AI felt productive. Actually making something? That felt weirdly hard, like a mental roadblock I couldn't get past.
It wasn't always this way. I used to tinker and build small projects regularly. But somewhere along the line, that creative spark fizzled. The momentum died. Now, the mere thought of starting something simple felt heavy.
One day, Ameya, an old college friend messaged—he was visiting Pune. A PhD and expert of AI domain. I asked him, instead of traveling or spending time chatting, would he be interested in building a couple of AI projects together when he is in town. I thought by having a friend around, it would help me get past the barrier.
He replied, "It sounds fun, but I figured we'd mostly just chat."
"No talk, all action, just like old Startup Weekend days," I typed back, half trying to convince myself.
Of course, when we met, we spent hours catching up on our lives. By afternoon, I realized that we hadn't built a thing and the day was slipping away. Having known that this is the only opportunity to get past the mental roadblock, I said, "Let's go to Pagdandi, a cafe on the link road." Thinking a change of scene might help.
Inside the cafe, we ordered coffee and settled down. But the situation felt familiar. I was stuck staring at the blank screen. With the big question, 'Where do I even begin? What do I build? I don't even have an idea to work on.' Across the table, Ameya was already typing furiously, fingers flying across his keyboard, lost in his own project.
Ameya asked what I was working on, and I couldn't muster up any words. I wrote a bunch of ideas, but none resonated.
After scribbling for a while, two unrelated thoughts surfaced from the clutter. When I was running my previous startup, I went to an HR workshop, where we were divided into pairs and made to do roleplaying activity based on some situation. That workshop had an impact on me since I realized I am very impatient person, and I interrupt a lot while talking to others. Another was an AI chatbot tool (playlab) which I was now very familiar with. How about I combine these two? Can I make a chatbot where people can practice roleplaying?
Magically, for the next hour, the buzz of the cafe faded. It was as if I was into a higher dimension. Completely focused on making this little chatbot work. I put a simple prompt to get started. Tested the chatbot. It was extremely bad. I changed the prompt, and tested the bot again—the bot improved a little. After repeating this cycle for a couple of times, I was finally happy with the outcome.
"Try this," I said, sliding my laptop towards Ameya as evening approached. He played with it for a few minutes. "Kya baat hai Kabra. Not bad for an afternoon's work." I was happy that I built something after a long time. It wasn't fancy. It wasn't revolutionary. But I managed to break the long standing barrier.
Later that night, I was wondering, what shall I do next. I thought maybe the workshop organizer might find it interesting too. But I forgot the name of the organizer. Frantically, I started searching through my emails, looking for workshops I have attended that year. Finally I found her name but no email id. I opened up LinkedIn, found her profile, and opened up the inbox to send her the message. My fingers hovered over the send message. But, the internal debate started again. The chatbot is so silly and basic, why even bother her. I should just delete it and go to sleep. The urge to delete was strong.
What the heck, I thought. I actually built something today, and pushed past the massive inertia, even if it was just a bit for building a small chatbot. So, I hit send, "Built something today that reminded me of your roleplay exercises. Nothing fancy, thought you might find it interesting."
I closed the lid of my laptop, and went to sleep, expecting nothing in return. Three days later, I was on a video call with Protima. "I liked the chatbot, interesting work. But, the kids' version isn't what we need. Could you make something similar based on a case study? We are pitching to a client for carrying out corporate assessments, and your AI tool could be useful."
I was surprised. This little project that I had built, just to get myself unstuck, had sparked a real business need?
I replied, "I think I could try it out." A month later, Protima messaged "Good news! We landed the client. The client mentioned that among the three applicants, we were selected as we had an AI evaluation tool."
My small afternoon project, the one I almost didn't share, became my first paid AI work.
That cafe afternoon drove home a lesson: the real challenge often isn't the skills or the technology, it's overcoming the inertia to start.
Now, when I feel that familiar resistance, I remember the afternoon at the cafe. I schedule dedicated "building blocks" – non-negotiable time just to create something, anything. Even if it's something small as a chatbot. It's a reminder that getting started is often half the battle won.