I have a post sitting in my drafts titled “Death by Hype” which I haven’t published yet because it’s basically one giant depressed rant about how everything is going to shit. This post is obviously a different one, but it deals with the same overarching theme.
An Innocent Plugin
I’ve been using Obsidian for a little over a year now to keep track of ideas and not down notes for projects. Over this time, I’ve found myself increasingly using it to list to-dos and tasks as well, and found myself wishing for a good Kanban-style way to display and manage them.
There are a couple of plugins offering this functionality, but they either don’t quite do it the way I’d like them to, or are unmaintained. At least that used to be the case until I recently discovered a new one, which integrates with Obsidian’s new “Bases” feature.
At a glance, it looked awesome, like a perfect fit. Simple, maintained, with just the right amount of features. I installed it, played with it, enjoyed it. But as I browsed its GitHub repo, things started smelling funny. Something felt off.
Everything seemed so well-structured, but soulless. The project hadn’t been forked from the Obsidian base plugin like most others. There was an AGENTS.md file. My trust in the project tanked.
Was I using, perhaps even enjoying slopware?
No Hard Proof
Now, I don’t know how much of the code was written by an LLM, how thoroughly it was reviewed, cannot prove non-human involvement at all. This is just a feeling, a hunch—and a moral dilemma.
Is it right to use LLM generated software and tools while generally being opposed to the practice of using LLMs because of their ecological, ethical, economical and social implications? It’s a dilemma I’ve been dealing with for a while. It’s just so hard for me to draw a line in the sand in this brave new world.
Besides, this plugin isn’t the only piece of software I use that has very obvious LLM assistance. My static site generator of choice, Astro, is being developed in a machine assisted way as well.
They recently released their big version six update, which, for me at least, broke some fundamental things that weren’t documented as expected breaking changes. Again, I can’t prove anything, but there’s this sinking feeling in my stomach that perhaps generated code was not reviewed as thoroughly as human written code would’ve.
How long until basically every dependency I rely on has some form of LLM generated components? The whole reason I use dependencies is that it would be nearly impossible for me to build everything by myself. So just ditching them isn’t an option, unfortunately.
Coming to Terms
I feel like such a hypocrite. I don’t want to use these products and using libraries, plugins, tools built with these products makes me feel like I’m indirectly supporting them. And there’s more, I dabble with locally running LLMs. I’m guilty of shooting complex queries off to larger models from time to time because I’m curious what will come back. I’ve even asked simple queries when traditional search engines failed me, and I was pressed for time or energy.
It makes me feel shameful. Hollow. Like a fraud. These products are worming themselves into my life and avoiding them is beginning to feel like an impossible task I just don’t have the energy for. When I started working full-time, I chose to forego relying completely on free and open source operating systems and software for greater convenience and conformity to the larger system. It stung, but made things easier. I came to terms with it, eventually.
This feels similar, but worse. Probably because it seems like I have less agency in it, like it’s being forced upon me. Maybe I’m just more tired than I was half a decade ago.
I guess I use slopware now. It doesn’t feel good, but it’s the truth.