# Give living tissue and silicon separate, measurable jobs

Demonstrated components / proposed integration. Brainoware demonstrated organoid-based reservoir computing in a laboratory system. DishBrain studied dissociated neuronal cultures interacting with a simulated environment; a two-dimensional neuronal culture and a three-dimensional organoid are different model classes. These results motivate experiments on adaptive dynamics, not a claim that neural tissue is a drop-in CPU, a general intelligence, or a recovered mind.

Canonical page: https://givegigs.com/wiki/necro-chapter-xdpu-hybrid-architecture
Last updated: 2026-11-09T18:58:12.584Z
Parent topic: [necromancer-neuroscience](https://givegigs.com/wiki/necromancer-neuroscience)

## Contents
- [Overview](#wiki-section-overview-1)
- [Chapter notes](#wiki-section-chapter-notes-2)
- [Dossier section](#wiki-section-dossier-link-3)

<a id="wiki-section-overview-1"></a>
## Overview

Demonstrated components / proposed integration. This page expands the “Give living tissue and silicon separate, measurable jobs” chapter in the neuroscience collection.

<a id="wiki-section-chapter-notes-2"></a>
## Chapter notes

Brainoware demonstrated organoid-based reservoir computing in a laboratory system. DishBrain studied dissociated neuronal cultures interacting with a simulated environment; a two-dimensional neuronal culture and a three-dimensional organoid are different model classes. These results motivate experiments on adaptive dynamics, not a claim that neural tissue is a drop-in CPU, a general intelligence, or a recovered mind.

A proposed Necromancer stack assigns culture maintenance to qualified laboratory systems, stimulation and recording to calibrated interfaces, timestamping and interlocks to local electronics, and feature extraction and readout fitting to conventional software. A reservoir benchmark must separate changes in the living network from learning performed by the digital decoder. Report both so software progress cannot be silently credited to biology.

Proposed acceptance gate: reproduce one bounded temporal classification task on held-out biological preparations and compare it with an electronic reservoir and a simple digital baseline. Freeze the evaluation before testing, report unsuccessful cultures and setup time, and retain a software-only path if tissue adds no useful information. The long-horizon vision is a complementary biological instrument whose particular contribution has survived this comparison.

<a id="wiki-section-dossier-link-3"></a>
## Dossier section

[Open this exact section in the Necromancer dossier](/necromancer#xdpu-hybrid-architecture). The original chapter status is “Demonstrated components / proposed integration.”

## Sources

- [Brain organoid reservoir computing for artificial intelligence](https://www.nature.com/articles/s41928-023-01069-w)
- [In vitro neurons learn and exhibit sentience when embodied in a simulated game-world](https://pubmed.ncbi.nlm.nih.gov/36228614/)

---
This article is collaboratively maintained. Check the canonical page and cited sources for the latest context.