Pipi and its databases

Mike's Notes

It is time to write about Pipi 9 and why it has so many internal databases. I get a lot of questions about this.

Resources

• https://wiki.ajabbi.com/eng/9/engines.html

Pipi and its Databases

By: Mike Peters
05/01/2025

Mike is the inventor and architect of Pipi and the founder of Ajabbi.

Introduction

Pipi 9 consists of many autonomous agents, each a software program containing at least one relational database.

Example

Engines are a type of autonomous agent. The Namespace Engine is an example. It provides a unique name for every part of Pipi, ensuring message delivery, version history, security, and more.

• https://wiki.ajabbi.com/eng/9/namespace-engine.html

This namespace data is stored in a database part of the engine. There can be multiple copies of any engine, so the records in each database will be different based on where the copy of the engine is located.

In the Namespace Engine, there are 60+ tables.

Data Model

The databases inside Pipi 9 are designed by humans, built and populated by Pipi.

Evolution

Pipi 9 was designed to evolve and self-organise in response to its external environment.

These changes happen at the level of the autonomous agents and in their database records.

The database also stores instructional information on how any autonomous agent should behave. Combined with templates, this data can generate static code for later execution. As the data changes, new code is written.

Speed

Pipi 9 is slow because of its complex internal structure. It runs many batch processes and adapts slowly but is reliable, stable, and cheap. It is ideal for interacting with large, complex systems where minor course adjustments are often needed. It is more like a slime mould than an LLM—a super thermostat capable of slowly learning.

Emergent Properties

The result of all these interactions causes Pipi's emergent behaviour as a generative platform. I still don't fully understand how I got this to work—it was more accident than intent. Hopefully, as Pipi receives a front end, more light will shine on this question.

SaaS Applications

Q. What about the free applications available on GitHub in the future?

A. They will each use an industry-specific relational database model.

Inspiration

I got the idea when I read Markus Covert's 2014 article in Scientific American about successful computer simulations of Mycoplasma.

• https://www.blog.ajabbi.com/2023/12/markus-covert-and-mycoplasma.html
• https://www.scientificamerican.com/article/cell-model-the-inner-workings-of-a-cybernetic-cell/

Summary

The databases play a key role in this process because the records store the history of the path taken, phase changes, current configuration, settings, constraints, etc., which can change or "evolve" over time. This acts as a feedback loop as the engines and other autonomous agents interact with their environment.