Building the companies of tomorrow

We challenge the assumptions everyone else builds on — then use AI to build what they can't see yet.

Game 2 · Move 37

About

The move everyone called a mistake

In March 2016, DeepMind's AlphaGo played Move 37 in Game 2 against the Go world champion. The move was placed in a position that 3,000 years of human experience had deemed suboptimal — the fifthline.

Every commentator called it a mistake, but it ultimately won the game.

It was the moment a machine saw value where human instinct saw none.

Our Worldview

AI should not make yesterday’s company slightly better.

It should change the architecture of the company — how it observes, decides, acts, learns, and compounds.

  1. 01

    Essential businesses are underestimated

    The physical economy contains durable demand, scarce expertise, trusted relationships, and enormous workflow waste.

  2. 02

    Every workflow is redesignable

    We do not bolt an assistant onto an obsolete process. We ask what the process should be when intelligence is abundant.

  3. 03

    A company should learn from every transaction

    Traces, exceptions, outcomes, and feedback should make the next decision and the next agent better.

  4. 04

    Human excellence moves up the stack

    AI should remove coordination tax and administrative drag so people can apply judgment, relationships, craft, and responsibility.

  5. 05

    Systems compound

    Shared operating architecture, data, density, and M&A can turn individual improvements into a durable advantage.

  6. 37

    The non-obvious move matters

    Conventional wisdom is often a compressed history of old constraints. We look for what becomes possible when those constraints change.

Pioneers Wanted