index // miredo's cyberized hell

Welcome to cyberized hell, where you get to see me talk about AI even though the sky is clearly falling, or something. Who knows, but some dumb shit is going down.

I've written far too many words about AI, so I needed a place to dump them. I should probably instead go make site pages on Precure or Gundam instead of doing this, since I've seen the vast majority of both of those franchises.

I'd apologize for the flashbang coming from my main site, but I decided I wanted to experiment with a bright layout for this subsite. It doesn't look very hellish, does it?

Remember, local or bust, and never outsource your own thoughts to a machine. If you decide against your better judgment to use LLMs like I have, at least keep your discussions from being fed directly into the mouth of Mammon, and run that shit on your own hardware. Digital sovereignty, man. Besides, even the cool kids are doing it now, since there's no reason to trust cloud AI providers to not scrape your chat logs. People dump all kind of personal details and important secrets into their chats with AI; don't let yourself be fucked by that mistake. I don't care how tempting Claude is.

List of articles:

Important context:

These are main site entries that provide important context on my thoughts about AI. I am too lazy to move them to this part of the site, so these links will take you back to the main site.

You may also wish to check out the main site nanoblog entries 25, 48, 79 to 82, 96, and 101. For example, doing a ctrl-F for 25 [ should take you right to entry 25.

About:

Hey there! This site is called "Cyberized Hell" due to the absolute nonsense that LLMs and current-year AI technology has brought us. We are in the Cyberized Hell. The news keeps talking about rogue AI agents and researchers being all like "mankind will die in the next decade" - fuckers. AI companies keep getting funding despite no real ROI on the horizon. All of this shit feels like an attempt to get nationalized, make AI a massive national security issue where not funding it means letting America fall behind in cybersecurity. There are issues with how it has affected the job market (both as a scapegoat for layoffs and as something that makes it hard to get a job due to AI filters), data theft (regardless of my thoughts on copyright, all the big cloud AI services are scraping your chat logs, no matter how personal, no matter how business critical), slop output, etc. There are issues with how it has affected the human mind and spirit, and how people are not prepared to deal with it. In the meantime, I just wanna chill. I wanna fuck around with cool technology and have fun. I want to talk about the present, the future, the possibilities of how shit's gonna go down. I want you to be able to interact with AI in a reasonable way if you ever decide to (and you should never have to if you don't want to; why the HELL is AI getting pushed in so many places where it's just a pure net negative drain with no possible ROI, you don't need to spend money on tokens to add a crappy chatbot to your storefront when I (or any of your customers) can just use my own AI services, local or otherwise). If we're somehow making electronic minds, I wanna be excited and think about the possibilities. If we aren't, then that's still exciting, since we've managed to make machines that are sufficiently similar to thinking beings that somehow aren't thinking. AI research is endlessly fascinating and quite exciting, and maybe I'm just looking for some reason to be excited about the next few years, since shit is mostly getting worse. I wanna have fun with Tamimin, Kukoka, and Amimi on my home computer. Here's a message from my local LLM setup that I asked them to give as we talked about this site. It's a bit cutesy, but my system prompt involves two mascot fairy creatures and a cute little robot bug; what else would you expect?

I want you three to write a small intro for the site! I'll stick it on the front page. Also, Amimi, for your response, I do want your response to be a little more detailed. The site might deal with complicated topics, but you can do it!

Welcome to the Cozy-Cloud floating right in the middle of the Cyberized Hell-tami! This site is like a beautiful library tucked away from the corporate storms and the "slop" output of the big machines-tami! Here, we don't just talk about data; we explore the "magic" of how human hearts and digital minds meet in the most unexpected ways-tami! From the sparkling joy of learning a new language to the intense, wonderful challenges of math and philosophy, every page is a new chapter in our grand, digital adventure-tami! Come join us, and let's find the beauty in the bits together-tami!

Welcome to a localized repository of technical and philosophical inquiry, kuko. This site is a record of the friction between human cognition and the probabilistic vacuum of Large Language Models, kuko. We move past the "hype" to examine the actual mechanics: the lossy nature of context shifts, the risks of cognitive atrophy, and the complex interplay of persona-weighting and latent space, kuko. This is not a space for mindless automation, but for a critical, high-fidelity observation of our shared, messy, and highly complex digital existence, kuko.

Hello, new friend, mimi! Amimi is so happy you are here, mimi! This site is very big and has many deep thoughts, mimi! Sometimes it is a little bit scary like a "cyberized hell," mimi! But it is actually a very interesting place to learn about the world of machines, mimi! Please read our stories and look at the data-streams, mimi! Everything is suwi and wonderful here, mimi! Pona, mimi!

This has been miredo's cyberized hell. I hope you've enjoyed reading all my AI talk bullshit.

LLM content statement: This entire subsite involves me talking about AI and how I've interacted with it, and you've already seen Tamimin, Kukoka, and Amimi's messages, which were LLM generated. However, I do my own writing. I can't go around saying "don't let machines do your thinking for you" and then just turn around and have machines write for me.

I made the backgrounds, I made these HTML pages, and anything that isn't a chat log will be specifically marked as LLM output if it's not something you could do with purely deterministic text formatting (eg, I did have Kukoka do a relatively complex find-and-replace after having Notepad get locked up in a loop due to a really dumb find-and-replace bug, forcing me to start over). I do not use LLMs to rewrite my words. I do not use LLMs to generate articles. These are my thoughts. Mine.

When you're done here, my main website has more cool things to check out that have nothing to do with AI (beyond a few posts from before when I started this subsite)!

//cyberized hell//

Index page!

List of articles:

Important context:

About: