Great read. I always found it so ridiculous when people assert that expecting a modicum of tech literacy is “elitism”. No, it should be the bare minimum; it’s because I naturally assume you respect your own time and autonomy. It’s not even like it’s complex, you don’t need to be a CS grad to just… have a basic understanding of how a computer works and use it outside of a corporate consumer pipeline.
As I wrote elsewhere: the likes of Apple have been trying to turn people into unthinking consumers for decades, but the massive spike of anti-intellectualism in recent years has exacerbated the problem.
Awesome article, thanks for sharing.
Gives a lot for me to think about, how the spaces I grew up with tech have been eroding away and why in ways I’ve noticed but never really put together.
I’ve never been a big fan of the term “power user” and I think there’s a little too much disdain for the people who aren’t- I think it’s reasonable for people to just want tech that “just works”- hell I moved off of Android when I realized I wasn’t really getting anything out of the tinkering and wanted a phone that just worked. I could stick with the computer for the fun stuff. (and as the article discusses, Android thinks so too!)
But there really isn’t any reason a platform can’t be open and also be functional- well except for the money! It becomes an economies of scale issue too- if most people don’t have or want open platforms it’s going to be harder and harder to have them.
While I can see a future where everything’s closed down- it’s been slowing happening to desktop computers too, like macOS gatekeeper is leading homebrew to remove a bunch of apps Apps like LibreWolf and FreeTube are to be deprecated in 2026-09-01 · Homebrew · Discussion #6334 · GitHub - we don’t need to give into that. I don’t feel despair over this- I feel energized to do what I can to keep the spirit of tinkering on computers alive. Let’s keep building cool stuff.
I unfortunately wasn’t able to get through more than a few paragraphs of that post, as I suspect it has been, um … heavily “assisted.” If you catch my drift.
I was thinking the same thing @st3phvee - there was a lot about this that set off the LLM flags in my mind :\ Which like…not to be mean but isn’t that a little ironic given the topic at hand?
Anyway just speculating. I do think about this topic though, like I was helping my friend install something the other day on her computer and she seemed a little nervous like, worrying she was gonna get a virus or something just because i opened Terminal to do something! Everybody starts somewhere and has different priorities, so totally not judging her for that, but it did take me by surprise a little. She’s in her 30’s like me and hasn’t ever seen the command line stuff outside of movies/TV.
That kinda situation makes me wanna learn tech more and get deeper knowledge so I can help others along with myself I guess, it makes me a lil sad how…not optimistic it feels to think a lot of us depend on our tech but don’t know a lot about it
Glad it’s not just me! I forced myself to read the post in full and pasted it into my notes app so I could highlight the various tells… “It’s not just X. It’s Y” (and variations on that sort of phrasing) come up at least 24 times, sweet jeebus. 
I probably missed a few, as my eyes started glazing over towards the end…
I’ve spent a lot of time talking with people on Reddit who are open about their LLM use, and a lot of them would absolutely identify as “power users” of this tech. They prioritize their “ideas” over the words themselves, and consider the writing process secondary to the “end result.” It would definitely be ironic if someone lamenting the supposed loss of a tech culture that celebrates “technical competence” did indeed use an LLM to write this post, but the author probably wouldn’t see it that way. Heavy LLM users tend to see themselves as “directors” or “producers” who are in full control of the LLM’s output.
There are some ideas expressed in the article that I can agree with (once I turn off that part of my brain that recoils at “LLM-speak”), but then there’s stuff like this:
Ask a twenty-two-year-old to connect to a remote server via SSH. Ask them to explain what DNS is at a conceptual level. Ask them to tell you the difference between their router’s public IP and the local IP of their laptop. Ask them to open a terminal and list the contents of a directory. These are not advanced topics. Twenty years ago these were things you learned in the first week of any serious engagement with computers. Today they’re exotic knowledge that even a lot of working software developers don’t have, because you can go a long way in modern development without ever leaving the managed abstractions your platform provides.
This is a rather obvious example of the stochastic parrot being confidently incorrect about something. The “not advanced topics” described here absolutely were not “things you learned in the the first week of any serious engagement with computers” back in 2006. Come on. Only a small fraction of “serious” computer users would have been familiar with these things twenty years ago. I would actually argue that there are more people familiar with things like Terminal today, as Linux is more popular now than it was in 2006.
EDIT: Also, I still can’t get over this:
You’ve built a generation that can’t extract a zip file without a dedicated app and calls it innovation.
This is the stochastic parrot fabricating nonsense in an attempt to sound “sarcastic.” What “generation” are they talking about here? The generation that used WinZip and WinRAR? Y’know, the software that was built in 1991 and 1995 respectively? The “generation” being poked at throughout the article (the “average person who grew up with smartphones”) wasn’t even born yet when these “dedicated apps” were first invented…
Working as a sysadmin, something I find annoying is that tools are being abstracted even at a professional level because so much is being sold as a service rather than a product. Maybe it’s just within my own niche, but since we started relying more on cloud apps, a lot of the tools I admin are total black boxes. If something goes wrong with a cloud service, I submit a ticket to a rep who then sends it off to an engineer (so, someone is doing the tinkering and problem solving), whereas even just 10 years ago I might have been the one to, like, physically restart a server or push a patch. For a lot of cloud apps, there is no opportunity to become a power user even as the sysadmin. Cloud companies build and host tools, and we just pay them to keep stuff running. And if there’s a change we don’t like (like companies adding AI to EVERYTHING, ugh) there’s nothing we can really do… it reminds me of the dynamic of subscribing to Netflix instead of owning DVDs. It certainly makes my life easier, since so much is out of my control, but I wonder how much knowledge/learning I’m missing out on. Maybe it’s just rose tinted glasses though! It’s a lot easier to tell someone a fix is coming from a cloud company than having to actually troubleshoot the fix myself.
Anecdotally, I totally agree with you. I work with a wide age range of users (18-80+), and there doesn’t seem to be a knowledge bell curve compared with age like the article implies. Different users of different age ranges might be more or less comfortable with tech for different reasons, at least in my experience, and it can be a bit reductionist to categorize people by generation. Back in the late 90s and early 00s, some of my older coworkers were scared their IT jobs might be obsolete in the future once everyone was a “digital native” (remember that term?), but it seems that over time people are just as “bad” with tech as they ever were. The difference now, I think, is that 1) tech isn’t niche anymore and 2) we don’t own things as much as we used to, and rely more on cloud apps.
Yup. Being reductionist is one thing that large language models truly excel at, as they were trained on all the “back in my day, we could do XYZ” generation vs. generation tripe that gets posted to Reddit on the regular. I wouldn’t be surprised if the original ChatGPT prompt for the article looked something like this:
Bingo! I think the impression that there are increasingly fewer power users in computing today is mainly a consequence of scale. ChatGPT the author of this piece would do well to remember that home computing exploded at the end of the 90s and the turn of the millennium; households went from having maybe a single family computer in the living room to multiple computers for each family member. Of course it’ll seem like power users are a dying breed when we’re talking about something that has gone as mainstream as home computing! Blaming it on the greedy “technology industry” and “kids these days with their smartphones and their tablets who don’t know how anything works” is such an oversimplification that is also unfairly dismissive of an entire generation.
I’m a Millennial who grew up with computers, not smartphones, and I know for a fact that there are 22-year-old former “iPad kids” out there who could run circles around me on any computer. Being a “power user” of anything (not just computers) has absolutely nothing to do with any sort of perceived generational divide. There has always been a small coterie of power users and there will always be a small coterie of power users. People who love to tinker with stuff and learn everything there is to learn about stuff are (and always have been) a minority.
I do agree with you, though, that the push towards third party cloud services is an issue. I think we can also add increasing reliance on third party web development frameworks to the list as well. But the author’s central thesis that power users on the whole are dying off and no one cares… just… no.
Oh I was just about to make a comment about the terminal stuff too. They weren’t even teaching terminal stuff back in the late 90’s. (it’s real funny considering my gifted program had us running through QBasic activities, but almost nothing about basic navigation and terminal commands) And honestly, even with running a VPS I couldn’t open a terminal and do that off the top of my head without a command cheat sheet because my brain scrambles syntax. Like, being able to know how to troubleshoot what you don’t know accounts more for your ability to adapt to a platform that it does to remember the exact words and syntax for a command line.
Also agree it’s really odd to pick at WinZip/WinRAR. And correct me if I’m wrong, but I’m pretty sure Windows in the 90’s and early 00’s didn’t even have any native means of unpacking RAR and TAR files without those programs so like what exactly is the point of digging at those applications?
I don’t think the article was specifically referencing those programs. I mentioned them as examples of older programs that were conveniently forgotten about by the LLM in its effort to prove the prompter’s bias that “kids these days can’t do anything without a smartphone app.” This is one reason why using LLMs to write is so sketchy… they’re very good at generating text that sounds correct (especially when you’re just skimming it), but when you take a closer look, you realize that what it has written makes no sense and frequently contradicts itself. ChatGPT whining “you’ve built a generation that can’t extract a zip file without a dedicated app” is pretty dumb when you consider that most of the old school power users that are supposedly “dying” out used WinZip all the time back in the 90s! I remember when WinZip was one of the FIRST things you installed on a new system, because it was such a crucial piece of software! Also, the ability to easily extract compressed files has been native to all the major operating systems for many years now. Kids who grew up with smartphones aren’t using dedicated apps to extract compressed files, LOL. They’re just double clicking on them, like the rest of us…
Exactly. That whole paragraph in the article was just bizarre and not at all reflective of what it meant to be a “power user” twenty years ago (which, again, was 2006).
Edit: I see the article was also posted on Hacker News. Similar objections there: The Slow Death of the Power User | Hacker News (I usually disagree with most of the comments on Hacker News, so this was a pleasant surprise, haha.)
Initially, I passed over this thread until I popped in and saw everyone talking about this part
Ask a twenty-two-year-old to connect to a remote server via SSH. Ask them to explain what DNS is at a conceptual level. Ask them to tell you the difference between their router’s public IP and the local IP of their laptop. Ask them to open a terminal and list the contents of a directory. These are not advanced topics. Twenty years ago these were things you learned in the first week of any serious engagement with computers. Today they’re exotic knowledge that even a lot of working software developers don’t have, because you can go a long way in modern development without ever leaving the managed abstractions your platform provides.
and then I decided I had to give the article a read, because surely it would make more sense in context, but it really doesn’t. The examples given are so random. Why would a 22 y/o need to know how to remote into a server or list a directory in a terminal? Knowing what a DNS is, kind of makes more sense, but at the same time… It kind of feels like expecting a 22 y/o to know what zip codes mean. Yes, that could be interesting information and it’s information they use regularly, but knowing the meaning doesn’t really add or detract anything. The only thing that seems like it might be useful knowledge would be the difference between public and local IPs, but even then it’s not really something anyone’s going to use.
But also, all of the examples given are incredibly easy to search, and personally that’s the real distinction of a “power user” not their ability to recite random factoids, but to have the skills and knowledge to problem solve on their own.
I also found this part interesting
The early web had a “view source” ethos: you saw something interesting, you looked at how it was built, you learned from it, you made something of your own.
Because it reminds me of some of the “forum revival” discussions. Where view source does in fact still exist and you can still inspect websites to look at how the front end is built and learn from it. I do it all the time. The difficulty in understanding the frontend source of the examples given isn’t really a comment on the user, but on how the web has grown and become more complex than just a static web page.
This part also feels like it’s missing the point by a mile
Tutorials are not documentation. A tutorial teaches you to perform a specific sequence of steps to achieve a specific outcome. The steps are usually correct for the specific scenario the tutorial covers.
It takes a very black and white view of tutorials being bad and documentation being good, but it’s not a one or the other situation. The only way to truly learn is through a combination of various different things which includes tutorials and documentation. It comes back around to my earlier point. Tutorials do only give a specific sequence of steps, but that specific sequence can then become a foundation for a user to then expand their knowledge by having a better idea of what information to search for if they want to do x, y, or z.
Overall, an interesting article concept, but it doesn’t seem to understand what it’s actually talking about. I got whiplash when it randomly pivoted with
I want to be precise here because people get defensive fast.
The problem is not, primarily, that services collect data.
because I’m still not sure where that came from or how that’s the precise problem when the discussion up until that point was about users not knowing how to do things…
Agreed. And again, the reason (I strongly suspect) why this random stuff comes up in the article is because it’s LLM-generated, and LLMs don’t actually understand what they’re talking about… They’re just predicting the next most-likely word in a sentence based on their training data, and doing so in a way that sounds convincing until you closely scrutinize what’s actually being said. The subject of data collection often comes up in tech-related articles, so the LLM just threw in a random aside about data collection that makes no sense in the greater context of the article to meet whatever word count had been specified in the prompt.
I also laughed out loud when I got to this part. After spending approximately 4,000 words lamenting the slow death of the power user, it trots out this:
The power user isn’t dead. The skills exist. The communities exist — smaller, grayer, more scattered, fighting an institutional headwind that grows stronger every year. But they exist, and the knowledge is still propagating in the spaces the platforms haven’t fully colonized.
Then it goes on to say:
The obituary for the power user is being written right now.
Cool. 
(Ugh.)
I’m sorry to keep belabouring the point that this article is almost certainly AI-generated, but this stuff really irritates me. I usually hesitate to label stuff as definitely AI-generated (because there’s always a small chance that I’m wrong, and that the author’s writing style has simply been influenced by overuse of LLMs, or they’re Kenyan) … but this one feels pretty egregious. There are so many inconsistencies in the argument, so many random references to things that don’t make sense in context… people tend not to make mistakes like that (while simultaneously overusing every single grammatical LLM tell ad nauseum), but LLMs sure do!
I do agree, though, an interesting article concept overall. I’d like to read a human take on it. 
Huh. If it’s AI, that’s disappointing. I liked his initial series a while ago on Linux accessibility (or lack thereof), which was written in a similar style but with a lot more specificity. I can’t imagine LLMs have enough training data about the limitations of Linux screen readers to output writing about that, but maybe I’m wrong.
They have data on (edit: just about) everything that has ever been publicly written on the internet. Their training data includes everything in commoncrawl.org, just about every forum that has ever existed, every Reddit thread, every Stack Exchange post, every archived website in the Internet Archives, and so on, and so on. They do filter their training data, so it’s not LITERALLY everything, but if anyone has ever complained about Linux accessibility on the internet at any point in time in a knowledgeable way that meets their training standards, LLMs have that data in their training set and can regurgitate information about it in seconds. They scrape the internet for more data to regurgitate and extrapolate from every second of every day. 
Just for fun, I looked over all of fireborn’s other posts and they all read with the same style and diction. Judging by how niche the topics covered are, I have to assume that there were some notes/thoughts given to a genAI in order to output these (unnecessarily long) posts.
It really makes me wonder, to what end? I don’t really think people that do this see the forest for the trees. Like okay, sure, you only have to type a few notes and the chatbot outputs the article much quicker and with much less effort, but then what?
I’ll repeat myself in saying that I’d rather see a messy, unfinished human work than a fully-fleshed out AI generated one. Fireborn’s blog would be so much better off if he just published whatever notes he’s giving to an LLM directly to his site instead.
I’m glad you brought this up, because I also found this point to be odd. Maybe its because I’m not a “power user” or whatever, but I’ve always found tutorials way more helpful that vague general documents. Concrete examples help me understand how something is applied, how it can be used. The idea that everyone who uses a tutorial is just copying and learns nothing is laughable. Humans learn by experience and apply what we learned to new scenarios all the time. We can easily take something specific from a tutorial and then use what we’ve learned to figure out something a little different.
I noticed this too and thought it was weird. I have to say when I first read this article I didn’t have suspicions that it was AI, (most of what I read online are blog posts from people in this forum, so I have little exposure to AI essays). But yeah, after all the inconsistencies everyone on this thread pointed out, it makes sense. Ironic considering the same essay criticized LLM use (possibly why I didn’t suspect it):
LLMs have accelerated this to a degree that should make anyone who cares about software quality genuinely alarmed. You can now write complete programs without understanding what a single line of them does
Apparently the author does not have the same feelings about writing as they do code.
Hm, I don’t mean to insult Fireborn, or make a judgment on their character/morals. I’m not someone that sees genAI users as “bad” (its the systems that are), rather I see it as a detriment to fireborn themselves, first and foremost.
Anyways, I wanted to actually get into the discourse of the actual topic at-hand. I feel as a Zillienial/cusp baby that both people older and younger than me are less technologically literate.
Consumer-grade technology (especially PCs) were much more unstable, error-prone and easier to pry the hood off of when I got into computers. You had to modify the registry or find the %appdata% folder or replace .DLLs when bullshit broke.
I also was poor growing up, and my PC in high school couldn’t install Windows and I was forced to switch to Linux (specifically CrunchBang #!, RIP) which looking back, changed the trajectory of my life and is probably the main reason I’m a power-user today.
Yeah, no disrespect intended at all to the person behind the blog. And again, I’m sorry to have belaboured the point. LLM-generated anything just irks me so much; my problem is always with the LLMs themselves and the techbros who released them with little regard for the damage they’d cause, not the ordinary people who use them.
That’s a great point you raise re: the instability of older systems. I was the person my family members and friends turned to when something went wrong with their Windows computers, because I’d spent so much time troubleshooting mine (dear god, the Windows ME days…). I wouldn’t have called myself a power user back then, but I did have a lot more knowledge than the average person using computers in my personal circle. I think that’s mostly because I generally enjoy tinkering with stuff and fixing problems, but also because it was sort of a necessity back then if you just wanted your shit to work.
I hardly ever get family or friend requests to fix computer issues anymore. People haven’t really gotten any less tech-illiterate overall than they were in the 90s and early 00s. Their systems are just quite a bit more stable now (if they’re even still using laptops / desktops at all).