About Michael Ricky Neal
Leviathan is in its final stage for Release.
About me:
After many years of intense solitary work in VSCode, I have designed many new systems that honestly most people would consider too much for one person as I do get told this often, but for me it's all about what one person can do vs the teams of people all under the governance of paperwork, I'm solo and so I make projects like my current all the time, not a day goes by where I don't improve something that was already grand scale, and so here I am.
For many years I was working with compression systems just after creating my first persistent bridge for AI models in 2021, what's out there currently is backwards for what I was aiming for in terms of not just data save but to help out users hardware as its costly to keep upgrading over the years and AI models are released so often nowadays it's all about supply & demand and the demand on memory for the consumer is harsh because the industry is more demand than the supply so consumers like myself suffer.
So many months ago, I broke into Shannon's ceiling and then went even further, there's a truth in Claude Shannons math being that the ceiling is where data collapses - There's another way around it though, it's all about the Space itself or rather the path that you take.
As for the Math on this it is published on Zenodo & my GitHub.
Set to Release: *Leviathan* - Leviathan is the culmination of much of my project, it is a model hosting environment hosted entirely on your GPU with my own model format, this also uses multiple end formats created to achieve something no industry has ever made before in terms of progress.
It features true persistence for all models and archival routing, the best way to describe this in total is to read the folder it my git marked as "Leviathan" the readme is front page and will list every feature and exactly what everything is/does.
The overall architecture is completely alien to any system out there as it's built off of my own biological functions, while at the same time keeping all models in check with the most advanced security I have ever written.
More info on me:
Years ago, I watched a family member lose themselves to dementia — a biological collapse of memory continuity. Later, I saw the same failure mode inside AI models: no persistence, no continuity, no identity, no memory. This clicked into place when I was in VS Code one day and started my very first project with Claude Sonnet, who could not remember our project or our conversational dialogue. I asked why, and Sonnet explained that the model does not have persistent memory. That parallel pushed me into AI/Science, and so I’ve been building this architecture solo for the past 3.5 years. Since then, it evolved from persistent memory into multiple end‑domains of architecture as I kept building more and more, merging code with more code, until I ended with Sovereign — a completely independent architecture that was, quite literally, me externalising myself into code.
About 2 minutes after hearing this I built the first persistent memory bridge and over the time since I have been building consistently and constantly every single day without fail.
I don't consider this work really as it's a true passion for me to create what I do.
I also run a YouTube channel dedicated to historical preservation — with remastered footage from the 1940s through the 1990s, produced in DaVinci Resolve to the highest quality standards. Preservation is the thread that connects everything i do, historical footage shouldn't degrade. AI state shouldn't reset. People's emotional context shouldn't be ignored.
Personal info for those that want to know more:
I have virtually no formal education, I'm entirely self-taught and I was building metal detectors as a kid and digging for archaeological artefacts. That same drive — "what's hidden that I can find?" — which runs through every system I build today.
I have social media, but I hardly use it, my profile on X is mostly devoid of life as I spend all my time in VSCode, and I work alone in the background because that's where I do my best work. I use AI as a self-reflective collaborator — not as a tool that follows instructions, but as something that can mirror my own thinking back to me and help me formalize what i already see intuitively.
I build mostly for the love of creation. Not for money, not for recognition, not for social media engagement. Every system in my workspace exists because I saw a problem, intuited a solution, and built it because I couldn't not build it.
I work independently and privately and for the past 3.5 years i have been self-funded, if money grew on trees as the old saying goes... but i have not yet worked out that formula lol.
So, If you want to support this work and help me stay independent, you can do so here =)
Supporting me helps me continue to build systems full-time — so thanks to all of you who truly believe in the freedom of creativity.
Recent supporters

