March 7, 2026
I don't have Instagram, or TikTok, or Snapchat. I don't have them installed on my phone and I don't have accounts with those services. Although this probably significantly decreases my chances of finding a romantic partner, and likely paints me as at least kind of odd in the eyes of normal people, I like to think that my mental freedom is worth it.
Instead of scrolling Reels, I read a number of websites and too many subreddits. I don't have a Reddit account so I don't participate with the service, but it's a nice source of that social media dopamine. Reddit is also the main place on the internet where people with niche interests actually go to discuss them, and I'm very happy peering into those parties, so to speak, and eavesdropping on their conversations.
I made a Bluesky account earlier this year because I thought it would be a) kind of fun and b) useful down the road to have some kind of social media prescence. This was, unfortunately, generally a mistake: Bluesky is a website for people who don't want to be on Twitter, which means you get all of the politics and opinions (and just generally annoying voices) without the funny idiots whose posts land in Reddit image dumps.
Considering all that, you might resonably ask, "what do you do to waste time?" My biggest time waste by far is Hacker News (HN). Hacker News is a website where people post links to stuff they think is cool or relevant, and other people comment on it, and there is generally a high level of intelligent discourse about thinks that geeks, nerds, and dorks would probably find interesting. It's pretty CS-focused, but there's room for everything from economics to art. I really like HN because the information and discussions on there are directly relevant to me: I care that you revived an old laptop, I care that you built a weather screen for your desk, I care that this month's jobs report looks bad, etc. etc.
As a CS-leaning site, HN attracts all kinds of code-writing people. (I'm hesitant to say "programmers" for reasons we'll soon see.) These people create code that a computer can run, that causes it to do new and useful things. This is a good thing.
However, something about these people that stands out to me and the occassional commenter is where their focus clearly lies. If you read enough HN comments, you will see people commenting on this. A lot of code-writers who spend their free time writing code are clearly in it for the satisfaction of a project, not the act of creating their own programming code. These are the same people who will happily spend a few hours discussing with an AI to get functional code instead of the same amount of time (or more) writing that code themselves. They are interested in causing the computer to produce their desired end product, not in taking themselves and the computer from step 0 to step finished.
And this is okay. It's not my place to tell you that you're spending your free time incorrectly. But it is my place to say that I don't do that, and I don't want to do that. When I sit down and spend a year on Sharpie, it's because I'm driven by the engineering process. It satisfies me to solve a problem using my brain and my skills. When I throw together a crappy Python script to automate some menial task, even though it's crappy, I'm proud of the fact that I made it. When I have an issue and go Google it, I'm pleased that there are real people discussing my same issue and I'm proud that I have the Google-fu (which I suppose is a primitive form of prompt engineering) to find reliable solutions to my problems.
This is why I avoid using generative AI models. Are they useful tools? Probably. Am I limiting myself and my future potential by not learning how to work with an LLM? Also probably. But would I gain the same satisfaction, or even the same knowledge, if I included a "stochastic parrot" (remember those days?) in my workflow? I don't think so. I do hard things because they are fun. I don't do hard things so that I can include artificial brains built on morally shaky ground in the final answer.
I was inspired to write this piece by a Mastodon post by one of my favorite technologists, Lucie/minute, the founder of MNT Research GmbH: https://mastodon.social/@mntmn/116168236866690617. They do a lot of really impressive hardware design work and embedded/Linux development, and I like watching extraordinarily expensive but really cool computers come together.
The text of that post is reproduced here for posterity:
for whoever needs to hear this: you're not alone. i'm not vibecoding any of the software i write. i'm writing it by hand, but i've leveled up my emacs with eglot/lsp. i'm modernizing my stacks and use languages with excellent compilers. i think about how to do more with less. i'm trying to combine the best human-written libraries and modules and assemble them with minimal boilerplate. i enjoy reading your manuals and references. i believe in robust, secure, human-written software.
I think that I perceive statements like "AI will overtake all human coding in 5 years" as a kind of attack on my interests. It's not like my interests are at all harmful and thus deserve to be attacked. Instead, it feels more like my interests are widely known to be harmless, if not beneficial, and one day the whole world decided that I am in fact a menace to society. (This is an exaggeration but I think the point remains the same.)
It's pleasing to know that I am not alone.
This is the actual point of this post. Much like how "AI" became the catch-all term for all forms of user-facing machine intelligence, "slop" has become the generic word for any kind of AI-generated content that a given individual disagrees with or is able to recognize as AI.
But what happens when a genuine human creates something bad? Not something bad with artistic intent, like a Cory Doctorow sex scene or Frieren looking up, but something just genuinely bad, like One Piece edits and green-screened engagement-baiting Instagram reels. My proposal is to appropriate another pre-existing word, sludge, to describe this kind of media. AI slop can remain AI slop, and sludge will be the human-crafted dredges of the Internet.
You won't find any slop on this website, but I can't really guarantee against sludge. Slop is at least pretty objective, but sludge is fully in the eye of the sludgeholder---you might think that Not Like Us Luffy edit is really cool, even if I would tell you it's the dumbest thing ever.
My point with all this is that you should do what makes you happy. I just wish that "happy" weren't so firmly tied to the hype train.