The path ahead

From an episode
to a trajectory.

Our roadmap is a sequence of research responsibilities, not a countdown to consciousness or a promise that every architectural idea will ship.

First: complete the consequential episode

A meaningful unit of progress begins with a concern, continues through a bounded action, meets an observable result, and returns to a later decision. The episode must retain enough context to explain why continuation, revision, or refusal was appropriate.

This focus prevents an expanding collection of features from being mistaken for integrated development. A tool, a memory store, and a self-description can all work separately while failing to produce the consequential organization we intend.

Then: test transfer rather than familiarity

A mechanism should be examined outside its most flattering demonstration. Does it preserve appropriate attribution when labels change? Does it help with a new task? Does it recover after an interruption without repeating an uncertain effect?

The comparison should include a simpler account, equivalent access to relevant information, and controls for added compute or context. These are proposed research criteria, not results already established across the whole program.

Build a public record people can use

We want the website to make the work legible through essays, dated notes, educational models, and bounded findings. Public explanation should invite a concrete criticism rather than demand belief in a private body of impressive-sounding work.

Reproducible public research packages require their own release review. Private repository access and sensitive histories will not be inferred from enthusiasm, nor will an inaccessible link be presented as public verification.

Beyond the screen

The longer horizon includes richer environments, creative practice, and carefully evaluated physical interfaces. Robotics, shared worlds, and continuing collaboration become meaningful when they add new kinds of encounter and consequence—not merely a new skin over a conversation.

Eternities supplies the wider company setting. Lunari pursues personal and creative computing. Neurasoft keeps the scientific and developmental questions visible. Progress in one can help the others without requiring them to claim the same results.

Public editorial synthesis · 16 September 2026. Internal source observations are dated separately; this is not a live operational report. See the reading room and publication standards for source classes.

To carry with you

Build the next step well enough that the step after it can surprise us.

Follow the question.

Explore the collection

The next chapter is not written

More than a moment.
A future worth having.

Understanding minds. Preserving possibility. Making room for what comes next.

Read our thesis

Explore research, Luna, Psyche Lab, and the journal.

Local site search · No query leaves this site