Inspiration

This is an active project that I have self funded under the Linux Foundation, it was originally started 2 years ago at Cloud Native Computing Foundation, when I proposed an AI operating system, and Brian Behlendorf encouraged me to do it and self fund it.

What it does

There is a virtual desktop,

A decentralized virtual filesystem layer, with connectors to multiple remote filesystem backends, adaptive replacement cache, write ahead logs, and graphrag indexing

A decentralized decentralized dataset query, manipulation, ETL library, with a common crawl search engine, a set of formal logic theorum provers custom made to handle legal reasoning, and a decentralized knowledge graph database, and ability to convert any document into a knowledge graph and theorum prover set.

A decentralized model server, containerization, and agent supervisor, and scheduler to handle the bulk computation needed to perform tasks.

How we built it

Blood sweat and tears

Challenges we ran into

2 years of challenges building an entire operating system and machine learning architecture from scratch

Accomplishments that we're proud of

I am converting all of the legal corpus to formal logic, to make all of the law formally computable, and make an AI lawyer for everyone to have access to for free

What we learned

Way too much about formal mathematics, machine learning architectures, and the legal corpus.

What's next for Neurosymbolic AI Operating System + MCP++ protocol

Get it into a state where it can be maintained by the linux foundation, so that I can work on the business side of computational law.

Built With

  • openai
Share this project:

Updates