In October 2025, I was deep into building a sales engagement platform. Two hundred customer development calls with sales leaders. Every single one said they’d pay for it. The product worked. The problem was me.
I didn’t own the data. No matter how hard I convinced myself, it felt spammy. And the truth I kept dodging: I’m an engineer, not a salesperson.
Then an accidental invite changed everything.
The moment
A Carta-hosted tech week event. I watched a 22-year-old use Claude to build something live. Then the screen flashed a message: “Compacting conversation.”
It hit me like a gut punch. I was back in 1993, watching MS Access compact a database after a querying session. Thirty-two years later, the most advanced AI on the planet was doing the same thing — throwing away context to stay within limits.
In 2025, this should not be happening.
That thought wouldn’t leave. I needed to understand why. Not just the technical constraint — the architectural assumption underneath it. Why does every AI tool treat every session as a blank slate? Why does no one remember?
India
I flew to India to talk to Ragesh Krishna, former VP of Engineering at Shippable. If anyone could pressure-test this idea, it was him.
Two days after I arrived, my father passed away.
He was a traditional man. I had watched him perform the full rites for his own parents — every ritual, every day, for thirteen days. Hindu tradition calls for a period of quarantine. Minimal contact with the outside world. I decided to honour that completely.
Those thirteen days were the hardest of my life. But grief has a strange clarity to it. With little else to occupy my mind, I kept returning to the same problem. Why does AI forget?
The question that wouldn’t go away
When I emerged, Ragesh and I debated for hours. Days. The conversations were relentless.
His challenge was pointed: “If this is so obvious, why hasn’t anyone at Anthropic, Cursor, or OpenAI figured it out?”
It was the most irritating question anyone could ask. It was also exactly the motivation I needed.
We decided to try. Not to build another coding tool. Not to build a better prompt. To build the thing that was missing underneath all of it — a persistent intelligence layer that learns from every interaction and never forgets.
Athena was born. Our organizational intelligence engine.
Building Quarterback with Quarterback
We started brute-forcing the problem and quickly realised we needed to get fundamentally better at it. Along the way, people shaped our thinking — Devashish Meena, Ariel Tsetlin from Scale, among others — pushing us in directions we wouldn’t have found alone.
But the real inflection point came when we stopped using Claude Code and started using Quarterback to build Quarterback.
November 27th. The first time I launched Quarterback end-to-end. It was surreal. Like a toddler figuring out cutlery, or an adult trying to look cool with chopsticks — we stumbled constantly. But there was only one rule:
We should never compact.
That single constraint drove every architectural decision. Every experiment. Sometimes counterintuitive design choices that felt wrong at the time but turned out to be exactly right. Each session made the system smarter. Each failure taught it something permanent.
What we built
Quarterback is the world’s first organizational intelligence platform. Athena, the OI engine underneath it, captures decisions, patterns, conventions, and context from every session — automatically, permanently.
Our coding tool is not better than what’s out there. It’s on par. You’ll notice a lot of familiar functionality. Actually, you’ll notice what’s missing. No slash commands. No agents. No sub-agents. No Claude.md. No cursor rules.
That’s the point.
When your platform has real intelligence — when it genuinely understands your organization, your architecture, your decisions — you don’t need any of that scaffolding. The tool disappears. The work just happens.
We have a series of twelve posts coming that walk through the critical forks in our journey — the architectural bets, the failures, and the decisions that got us here.
With this, we introduce Quarterback to the world.