I don't understand why some people insist on recommending NixOS as a first Linux distro. It's so far from the norm and beginner friendliness they must surely be joking.

My motivation for using NixOS is maintenance. I've been running 2-3 personal Linux computers for the last decade, with one of them being a server.

To get stuff like services working with each other you sometimes need to make small changes to the config files of some services. The issue for me especially with the many services running on the server is coming back to a broken/misbehaving machine after 4+ months and now having to research what changes I made long ago and where those config files are buried.

Making the change and testing it would likely take less than 5 minutes if you had every detail you need fresh on your mind.

I simply don't have the mental capacity to remember all that stuff after months of working on other things. Especially if you're coming back to something broken this is a really annoying position to be in.

You want the fix now but have to start by looking up docs and trying to figure out what past-self did to get you into this mess, or to find out what has changed since then.

At some point I had enough and was either going to teach myself some sort of personal changelog / documentation system, or learn a new declarative configuration system.

Huge respect for anyone who can keep all this info in their mind and to those that meticulously update their own documentation, but I lack the discipline to do so in the heat of battle and will easily miss things.

Since then any system that I will have to maintain myself has been using some form of declarative management. It keeps all configuration accessible and organized in one place, so I don't have to go digging for the correct file paths. It self updates so that even when I go back and forth during testing I won't miss updating my standalone docs.

And NixOS brings this to my whole system. No old programs lying around because you forgot to uninstall them and have now forgotten about it. Same thing with pinned package versions that then wreak havoc once they're incompatible with the updated rest of the system. It even configures my goto tools (shell, editors, etc) to my personal liking when I set up a new machine.

Its not the first declarative system and probably won't be the last one I will use, but for now it really makes my life noticeably easier.

The character swapping really isn't accomplishing much.

• Speaking from experience, if I'm finetuning an LLM Lora or something, bigger models will 'understand' the character swaps anyway, just like they abstract different languages into semantic meaning. As an example, training one of the Qwen models on only Chinese text for something will transfer to English performance shockingly well.

• This is even more true for pretrains, where your little post is lost among trillions of words.

• If it's a problem, I can just swap words out in the tokenizer. Or add 'oþer' or even individual characters to the banned strings list.

• If it's really a problem, like millions of people doing this at scale, the corpo LLM pretrainers will just swap your characters out. It's trivial to do.

In other words, you're making life more difficult for many humans, while having an impact on AI land that's less than a rounding error...

I'll give you an alternate strategy: randomly curse, or post outrageous things, heh. Be politically incorrect. Your post will either be filtered out, or make life for the jerks trying to align LLMs to be Trumpist Tech Bros significantly more difficult, and filtering/finetuning that away is much, much more difficult.

i am aware of the thorn. i am also aware it is not used anymore.

thank you for the links, however.

his using alternate characters in an attempt to "foil" an LLM scraper is entirely a performative gimmick.

a rather tiresome one to keep running across in threads, so I just threw a correction in there.