# My Brainstem RNS-AI project has made progress

**URL:** <https://discourse.numenta.org/t/my-brainstem-rns-ai-project-has-made-progress/12627>\
**Category:** Science\
**Tags:** community, neuroscience\
**Created:** [September 28, 2026, 11:51am UTC](https://discourse.numenta.org/t/my-brainstem-rns-ai-project-has-made-progress/12627 "2026-09-28T11:51:53Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![unikum-sol](https://yyz2.discourse-cdn.com/flex030/user_avatar/discourse.numenta.org/unikum-sol/32/6710_2.png) [@unikum-sol](https://discourse.numenta.org/u/unikum-sol)\
**Post date:** [September 28, 2026, 11:51am UTC](https://discourse.numenta.org/t/my-brainstem-rns-ai-project-has-made-progress/12627/1 "2026-09-28T11:51:53Z")

</div>

Here is my current project status, it has become quite extensive in the meantime.

> **[GitHub - unikum-sol/brainstem: BrainStem is a biologically inspired, Real...](https://github.com/unikum-sol/brainstem)**
>
> BrainStem is a biologically inspired, Real Neuro-Symbolic (RNS-AI) cognitive architecture for lifelong learning. It is designed to learn models of the structures and dynamics of language and text through context hypotheses, uncertainty, contradiction, revision, neuromodulation, replay, and consolidation rather than by merely storing isolated facts.

I look forward to feedback and discussions.

Here is a small excerpt.

„BrainStem is a self-learning language-understanding system designed to acquire sentence-, word-, and relation-level structure from an unsegmented text corpus purely through statistical observation, without word lists, grammars, or filters. The system’s stated design principle is that every unit of knowledge, a sentence-level hypothesis, a word boundary, a relation between two entities, a category, or a question, begins as a **low-confidence, fully correctable hypothesis** , and only becomes a durable fact, relation, category, or question after surviving a multi-cycle, neuromodulator-gated consolidation process modeled on biological sleep-dependent memory consolidation.”

---

<div class="post-metadata">

**Author:** ![unikum-sol](https://yyz2.discourse-cdn.com/flex030/user_avatar/discourse.numenta.org/unikum-sol/32/6710_2.png) [@unikum-sol](https://discourse.numenta.org/u/unikum-sol)\
**Post date:** [November 16, 2026, 12:42pm UTC](https://discourse.numenta.org/t/my-brainstem-rns-ai-project-has-made-progress/12627/2 "2026-11-16T12:42:05Z")

</div>

Here is the latest status. If you are interested, feel free to take a look.

> <https://github.com/unikum-sol/brainstem/blob/main/docs/Project_Status_2026-11-16_2219.md>

Here is a small automatic summary.

„BrainStem reads large amounts of text and slowly builds its own understanding from it. It discovers words sentences and connections between ideas without being told the rules in advance. It forms categories groups related things and notices gaps in what it knows then turns those gaps into questions that guide further reading.

It can correct its own errors when new evidence shows that earlier conclusions were wrong. During quiet phases it replays and strengthens the most useful knowledge while letting weaker material fade. Over long runs it accumulates thousands of stable facts and relations that all trace back to the original text.

The system runs on its own for thousands of cycles grows its knowledge steadily and stays stable. Users can watch the process through an admin interface.”
