periman2 / philosophos

The project "Philosophos" is an AI research agent that engages in a perpetual cycle of research based on predefined goals.

AI Research Agent - Philosophos

This project was created as part of the lablab.ai hackathlon on AI agents.

AI Models Used (Subject to Change)

• GPT: OpenAI - gpt-3.5-turbo-16k
• Embeddings: OpenAI - text-embedding-ada-002

Overview

This is an always-on AI research agent built using OpenAI models and embeddings with the help of the langchain framework. It consists of a front-end website that showcases the "published" research and a backend service that hosts the AI agent itself.

Functionality

Philosophos engages in an endless cycle of research based on the latest set goal, ensuring a constant stream of valuable insights. This continuous generation of knowledge can provide any business or researcher with a significant competitive edge in any domain. The quality of research it is able to create is only limited by the quality of the AI models used and the data store it has access to.

Data store

For Philosophos to work as expected, we needed to feed its data store with a big amount of information relevant to the topics we wish to research. For the purposes of this project I have chosen the topic of philosophy and have injected its database store with books from the open and copyright-free choices of the gutendex website: https://gutendex.com/

Database

The project depends on a private Supabase database. If you wish to contribute and need access credentials, please feel free to contact me.