De Substratis Emergentibus Neuro-Network Substrates for Emerging Agentic AI Systems

Fuente: Zenodo
Gespeichert in:
Bibliographische Detailangaben
1. Verfasser: Medina Hernandez, Alfredo
Format: Recurso digital
Veröffentlicht: Zenodo 2026
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866901735000244224
author Medina Hernandez, Alfredo
author_facet Medina Hernandez, Alfredo
contents <p>Artificial intelligence is moving from isolated answer generation toward situated action. A modern AI<br>system may respond in language, call tools, inspect files, write code, execute commands, search memory,<br>ask for approval, launch tests, produce artifacts, and preserve evidence. The system is no longer only a<br>model. It is a network of possible actions.<br>This shift changes the problem of AI architecture. If a system has many possible paths, intelligence<br>is not only what it says. Intelligence is also how it chooses where to route attention, which tools to<br>activate, when to inhibit action, when to preserve memory, when to ask for human approval, and when<br>to treat an output as evidence.<br>This paper calls that underlying structure a <strong>neuro-network substrate</strong>. The term is intentionally<br>hybrid. “Neuro” points to activation, inhibition, trace, reflex, deliberation, and correction. “Network”<br>points to routes, surfaces, tools, memories, workers, registries, artifacts, and feedback loops. The term<br>does not claim that engineered AI systems reproduce biological brains. Instead, it provides a practical<br>vocabulary for systems that behave less like single functions and more like living operational networks</p>
format Recurso digital
id zenodo_https___doi_org_10_5281_zenodo_20247604
institution Zenodo
language
publishDate 2026
publisher Zenodo
record_format zenodo
spellingShingle De Substratis Emergentibus Neuro-Network Substrates for Emerging Agentic AI Systems
Medina Hernandez, Alfredo
Cognitive neuroscience
Artificial intelligence
<p>Artificial intelligence is moving from isolated answer generation toward situated action. A modern AI<br>system may respond in language, call tools, inspect files, write code, execute commands, search memory,<br>ask for approval, launch tests, produce artifacts, and preserve evidence. The system is no longer only a<br>model. It is a network of possible actions.<br>This shift changes the problem of AI architecture. If a system has many possible paths, intelligence<br>is not only what it says. Intelligence is also how it chooses where to route attention, which tools to<br>activate, when to inhibit action, when to preserve memory, when to ask for human approval, and when<br>to treat an output as evidence.<br>This paper calls that underlying structure a <strong>neuro-network substrate</strong>. The term is intentionally<br>hybrid. “Neuro” points to activation, inhibition, trace, reflex, deliberation, and correction. “Network”<br>points to routes, surfaces, tools, memories, workers, registries, artifacts, and feedback loops. The term<br>does not claim that engineered AI systems reproduce biological brains. Instead, it provides a practical<br>vocabulary for systems that behave less like single functions and more like living operational networks</p>
title De Substratis Emergentibus Neuro-Network Substrates for Emerging Agentic AI Systems
topic Cognitive neuroscience
Artificial intelligence
url https://doi.org/10.5281/zenodo.20247604