Independent AI research

Intelligence,
still becoming.

The next frontier is intelligence that keeps learning.
We’re exploring the architectures that could make it possible.

Discover our research
New architectures. Continual learning. Less compute.
Dense tetrahedra and niho manō teeth circulate continuously through the interwoven Kāula rings, following one another like an ouroboros. A muted olive crest travels around the closed loop.KĀULA / NIHO MANŌCONTINUOUS TRANSFORMATION
Niho manō — Becoming
RESEARCH AT THE EDGE OF POSSIBILITYScroll to explore

01 / Research

The questions
we’re testing.

Each project starts from one question and a fair baseline. Results are published as notes, including the ones that fail.

Self-modification and continual learning

Models that change themselves, or other models, after training, and the measurements needed to tell whether that works.

02Design

Interpretability-guided editing

When a language model edits another model’s weights directly, which interpretability information makes its edits work?

In design, on Rewire’s harness.

03Idea

Self-edits and peer ring

Can models improve by editing their own weights, or each other’s in a ring, and does a peer edit better than self?

Question still being sharpened.

06Design

Branching memory

Can a memory model that reads another model’s activations decide what to store and recall better than retrieval over text?

In design.

Learning rules

Rules for updating a network’s weights, found by search instead of written by hand.

Architectures and efficiency

How a model is structured, and what it costs to run.

09Idea

HILO

Can a planner working inside a learned world model complete long-horizon goals better than flat baselines, at lower cost?

Question still being sharpened.

Pass and fail are the pre-registered verdicts, quoted from each test’s analysis script. Every result is in the technical notes.

Latest technical note · November 13, 2026Rewire: can an AI improve a model by editing its weights and structure?A benchmark for agents that improve a model by editing its weights and structure, scored against blind edits of the same size.Read the note

02 / How we think

Big questions.
Careful work.

01

Start with the mechanism.

Ask what a system is doing, why it works, and what would make it learn differently.

02

Let evidence set the pace.

Build controlled experiments. Compare against strong baselines. Treat a promising idea as a question to test.

03

Make the work useful.

Connect fundamental research to systems people can understand, reproduce, and eventually put to work.

A LONG-TERM VIEW OF INTELLIGENCEKĀULA RESEARCH

03 / Perspectives

Questions worth
staying with.

A few ideas shaping our research agenda.

04 / The company

Independent minds.
An open horizon.

Kāula is an independent AI research company working toward intelligence that can keep learning.

Our focus sits at the intersection of adaptive architectures, continual learning, and efficient computation. We’re building a research practice around fundamental questions and the experiments needed to answer them.

Kāula is Native Hawaiian founded. Our name, kāula, is Hawaiian for “seer” or “prophet.”

Explore our research agenda
Kāula / Research perspective

AN OPEN RESEARCH QUESTION