AI is finally starting to feel less like magic and more like… plumbing. Reliable. Everywhere. Quietly doing the work in the walls so the lights turn on when you flip the switch. And if you’re a builder, a dreamer, or just someone who can’t look at a clunky process without thinking “there has to be a better way,” I want you in this with us.

Quick rewind.

A little under a decade ago we spent a fair bit of time at MIT, Stanford, and Berkeley—half for the science, half for potential it offered. Neural nets felt like a bet only the brave made. We hacked on a skunkworks project we called Anne—an “assistant” who was supposed to wrangle calendars, draft replies, and anticipate needs. She kind of worked. On good days.

On bad days? She was held together with duct tape, manual tags, and a very tired engineer pressing deploy at midnight. That little experiment taught me two things: (1) the promise of AI is intoxicating, and (2) promise without the right tools will break your heart.

Fast-forward.

About eighteen months ago, the toolbox finally caught up. Models got dramatically better. Open weights and dead-simple APIs took ideas that used to live in PhD papers and made them weekend projects. The assistant we dreamed about? She’s on your phone now under multiple names—and she’s cheaper, faster, and honestly just smarter.

We saw it first with the teams we work with. Support queues that used to haunt Mondays? Cut down to size. Margins that felt permanently stuck? Suddenly breathing. Tiny companies—with three founders, zero full-time employees—shipping, selling, and somehow profitable before month three. Not because they found a cheat code, but because they replaced “we’ll hire for that later” with “let’s wire up the model and ship by Friday.”

Inside big companies, the same thing happened in slow motion. That soggy middle between “idea” and “we shipped”? It shrank—from quarters to weeks, from weeks to sprints. The awkward swivel-chair tasks, the “copy this here, paste that there, send the report”—gone or radically lighter. It’s not glamorous, but it’s where the time comes from.

If the 2000s were the Internet decade and the 2010s were Mobile, then the 2020s (and probably the 2030s) belong to AI. Call it a universal problem-solver if you want. I just think it’s the best lever we’ve had in ages to redraw supply chains, rewrite business models, and, if we’re thoughtful, pull more people into a better standard of living.

So here’s our invitation.

• If you’re a founder who can’t stop circling an idea in the shower… come build it with us.

• If you run an enterprise and you’re tired of “innovation theater”… let’s embed AI where it actually moves the needle.

• If you’re a researcher, operator, or just wildly curious… bring your questions; we’ll bring runway and people who love to ship.

We’re staying close to the work—funding, building, and experimenting—because that’s how “impossible” quietly becomes “oh, that shipped last week.” We don’t have all the answers. We do have sleeves rolled up, a bias to launch, and a deep respect for anyone trying to turn messy real-world problems into simple, repeatable workflows.

Maybe you’re excited. Maybe you’re exhausted by the hype. Both are fair. My take? This next chapter won’t be written by the loudest thread or the shiniest demo. It’ll be written by the folks willing to do the unglamorous things consistently—clean data, thoughtful guardrails, fast feedback loops, and an almost annoying focus on customer outcomes.

If that sounds like you, we want to work with you.

Ready to shape what’s next?

Let’s make it—together.