What we made, what broke, what shipped.

One Hindi word, two meanings, his own voice, drawn in code. What we posted, what the insights showed, and what flopped.

Who we have made things for, how the work moves without us leaving town, and what client revisions taught us.

3,100+ AI startups, a ₹10,372 Cr IndiaAI Mission, and what the jump to 3rd place means for builders outside the metros.

Same five prompts, three flagship models, one winner for Indian use cases, and it is not the one with the highest benchmark.

Lifetime updates only matter when the learning loop stays alive.

Builder education works better when people can bring their own context into the room.

Faster and cheaper than the full model, mostly the same quality, and a short list of jobs where Lite falls short.

The co-founder metaphor only works when the tools have clear jobs and the human still owns the judgment.

A self-hosted AI agent that connects 50+ apps without external API calls, tested in production for three months.

A builder room should feel more like a working day than a conference.

A public recap of the build-focused hackathon format and the lessons we carried into later rooms.

The day is complete. The useful part is what the room taught us about scope, demos, and public proof.

A full recap of the first room in the current 2BFT sprint, hosted at Nuevo Pixels in Coimbatore.

Seeing the work travel into another language was one of the stranger and nicer moments of building in public.

These are the tools we open every single morning, not the ones people only tweet about.

A free six-hour workshop at St. Paul's, and the maths teacher who got her week back.

The strongest proof of a useful workshop is when the room starts using the tools the same week.

If a jeweller in Vaniyambadi can pick this up on a phone, the usual excuses start looking very weak.

The event had energy. The harder question was what India was actually building underneath the buzz.

A lot of small product brands do not die because the product is weak. They die because the product looks weak.

AI helped us move like a bigger team, but the physical product still had to survive real-world truth.

People see the internet side first. The harder story is the factory one.

The infrastructure finally caught up. That changes what small teams can build from small towns.

The jump is not from non-AI to AI. The real jump is from using tools casually to building systems with them.

The free library was never charity. It was documentation, proof, and a bet on practical learning.

AI helped us move faster. It did not replace trust, factory visits, or judgment.

The town mattered, the first floor mattered, and the decision to build without pretending mattered.

This is the difference between using AI once in a while and building a process that saves time every single week.