Machine Intelligence
from First Principles

We seek to implement authentic machine intelligence by learning from nature's blueprints, enabling ecosystems that adapt, scale, and integrate seamlessly under any condition.

Abstract geometric visualization representing First Cognition's approach to building intelligent systems from first principles

About First Cognition,

We're building AI components that mirror natural intelligence—capable of evolving, adapting, gaining insight, and developing behaviours that become effortless over time. Matching nature's patterns makes systems efficient, intuitive, enduring, and scalable.

Our Approach,
From First Principles

Visual representation of building from foundations - geometric shapes forming a solid base structure

Start from the foundations

We begin by stripping away assumptions—rethinking intelligence from first principles instead of layering complexity onto old systems.

Abstract visualization of knowledge representation - interconnected nodes and data structures

Redefine knowledge
representation

Our current focus is on knowledge foundations. Natural intelligence relies on adaptive knowledge. We are remodelling how information is captured, structured, and used—the essence of how natural intelligence understands and evolves.

Visual representation of evolving decentralised systems - dynamic network patterns and organic growth structures

Build evolving,
decentralised models

Our vision is for technology ecosystems to become self-organising and naturally adaptive. We believe this requires an approach centred on information rather than applications. This will enable a new generation of sovereign, governable, intelligent agents—and much more.

Who We Are

A Fresh Approach

Our founder, Pete Chapman, is an industry veteran who was already building neural networks in the 1990s. After decades working closely with industry leaders and seeing the data challenge firsthand, Pete is convinced that a foundational shift is required for information modelling, aligning more closely with broader principles of natural intelligence. Amorfs is the first step in this journey.

First Cognition is privately funded, with headquarters in Sydney, Australia, supported by offshore development in Vietnam.

Vision Statement

Our Vision for the Future

We envision a world where technology systems engage seamlessly with the information you already curate and control. Instead of applications that gatekeep your information, we think everyone should own their own information to share as needed. Ring-fenced agents should enrich our lives and free us from unrewarding chores, securely and transparently.

In the enterprise context, individuals, teams, and departments create communities of interest that evolve around shared information, freeing them from email chains and clunky web interfaces, optimised for agent support, and releasing energy for higher-value activity.

In our vision, systems don't just process information—they become self-describing, carrying all the necessary context and flowing seamlessly across boundaries, languages, and domains.

Intelligence no longer emerges from brute-force computation, but from elegant, dynamic, nature-inspired learning.

This is where we're aiming, and we think you should be too.

Pete Chapman

Knowledge that evolves with changing context is the foundation for intelligence, and essential for truly sovereign agents. Our Amorfs format, with its underlying graph model, brings a fresh approach to data. It's friendly for humans and agents, but powerful enough to support complex intelligent systems.

Pete Chapman

Founder and CEO

Amorfs Knowledge

Secure self-describing knowledge cube with shield and embedded schema patterns

Friendly but Powerful

Amorfs is an experimental knowledge format that's friendly for humans and agents, but powerful enough to support complex intelligent systems.

Fragmented data shapes converging into a unified provenance-rich knowledge stream

Rich Context

Context is foundational to Amorfs knowledge, including provenance, change history, time frame, and confidence levels. It turns fragmented data into interoperable, context-rich knowledge that agents and LLMs can trust.

Nested knowledge structure with orbital rings representing governance and analytics

Beyond Graphs and Documents

More than a graph, Amorfs separates concepts from data, enabling support for multiple languages, domains, and perspectives, including encryption and ZK-style analytics.