Last week, I attended Y Combinator’s first Startup School India in Bangalore.
I was genuinely excited heading there - not just for the talks, but to meet people building outside the campus bubble and see what people are actually working on in the real world.
~2000 builders from across the country, all in one place.
I spent the entire day there attending sessions, talking to founders across different fields, especially in AI and robotics. The room was full of energy, and it’s hard not to feel driven when you’re surrounded by people actively building.
The idea that kept coming up
Across sessions with YC partners like Jared Friedman, Ankit Gupta, and Jon Xu, and in conversations with builders, one idea kept coming up:
Live in the future and build what matters.
What stood out wasn’t just the line itself, but how people actually apply it.
Building what matters isn’t about chasing what looks exciting right now. It’s about identifying problems that will continue to matter - problems people genuinely face and working on them long enough to solve them well. There’s leverage in working on something that compounds over time.
That idea connected well with something Harshil Mathur spoke about. Razorpay was a hard problem space, but it was a real one. If you’re building in a space where customers genuinely need a solution, the difficulty becomes an advantage - fewer people are willing to enter, and competition stays limited.
The better question isn’t “is this trendy?” but:
can I see myself working on this for the next few years?
Starting before you’re ready
One place it’s easy to get stuck is waiting for the perfect idea before starting.
One of the most consistent patterns I noticed - both from talks and conversations was that the first idea is almost never the one that works.
Until now, I found myself trying to think through ideas and arrive at something “right” before starting. But most builders I met do the opposite: start small, make it cheap to try, and figure things out through iteration.
The clarity comes after you start, not before.
Start small, Grow big
That way of thinking showed up in very concrete ways.
One conversation that stayed with me was with a founder working on autonomous driving in India. Instead of trying to solve the full complexity of Indian roads - traffic, unpredictability, potholes, they’re starting with warehouse-to-warehouse transport on highways, where conditions are more controlled.
It’s a narrower problem, but one that allows faster iteration and real progress.
Solve that well, then expand.
What founder stories made clear
The same pattern came through in founder stories.
Listening to Aadit Palicha talk about building Zepto, it was clear the journey didn’t start with a polished idea. It started with small experiments understanding local demand, noticing dissatisfaction, and gradually moving toward a more controlled delivery model.
Their first warehouse was literally a cofounder’s apartment. At one point, they hit ~10k orders a day, and that’s when things became real.
The key wasn’t getting it perfect early - it was testing, validating, and only then committing more seriously.
Uncertainty is always part of the process. As Lalit Keshre put it, startups always come with question marks - around product, growth, profitability. The goal isn’t to eliminate them all at once, but to reduce them step by step through building and iteration.
What also became clear is that strong builders are obsessed with the problem, not the solution and definitely not with having the “perfect idea.”
Building in AI and robotics
This felt especially relevant in AI and robotics.
It’s easy to get pulled into trends - flashy demos, popular approaches, things that look impressive online. But the stronger signal is to stay grounded in the problem and choose the approach that actually solves it, even if it isn’t the most hyped.
What changed for me
Coming out of the event, the shift for me is simple.
I spent a lot of time trying to find the “right” idea before starting. Now it feels clearer that the better approach is to:
- start small
- keep it cheap
- iterate fast
- and let clarity emerge through building
If there’s one thing I’d want to remember from the day:
Don’t wait for the perfect idea. Start with something you care about - and iterate.
There were ~2000 people in that room, but the common thread was simple:
people weren’t trying to look like founders—they were trying to become builders.
P.S. Got $25k in YC AI stack credits from the event—excited to use that to build and experiment more.