Well, what a beautiful amount of time it has been on this platform! I honestly can't imagine how much support everyone (including your cats) has given me! And, honestly, I can't thank you enough. This all started out as a "what if" idea and now it's evolved into this. That being said, I want to share a little of the history of why I started this project and how it got here.

### Why I started "NILKNARFGonzo" as an account
When I joined Hugging Face, I was already enjoying using Ollama to run silly models on silly hardware. However, my journey on this platform all started with a link. I was reading the README for a random model, and I noticed a link to Hugging Face. I realized that this is where all of the magic happens, and Ollama is just how you run quantizations of these models. I found out that essentially anyone can make any model here, so I set up a plan to make a model that fits on a standard 3.5" floppy disk. I am amazed by old technology, and I just really like how creative people made creative solutions to problems that are long solved today. Floppy disks only had 1.44 MB of storage (usually) and I found that quite small; so I took that as a challenge.

Faced with the problem of having little amounts of storage, I decided to make a GPT-2 based model that somehow fit on this decades-old storage medium. This meant I was going to have to make a model smaller than a language model can be logically speaking. This crazy idea ended up getting thousands of downloads, on an account less than a week old. And, sure, it outputted words (rarely), but none of them made sense. That's when I realized that I can make bigger models to make models that made more sense. I had joined right around the time that there was a huge boom of small language models, apparently.

### How I got here today
With great power comes great amounts of time wasted, as literally nobody says. I figured with more parameters and more dataset there would be greater "intelligence" (but let's not go down the rabbit hole of if machines are intelligent). So, I increased the parameter count to hopefully get better results, but all I got was the ability to become insane. After the success of my first two models, I thought making a third would be even more popular. As a surprise to absolutely nobody, that happened almost instantly. And I just want to thank everybody for helping support my project. Thank you.