Answering your question: Linux for gaming is kinda insane these days. With the exception of a small number of competitive multiplayer games where the devs are actively working to sabotage progress, pretty much anything runs as good or better on Linux now than it does on Windows. Even vintage games that Windows struggles to run.
Many other cases of things just improving over time to go from irredeemably broken to working so well you don't think about it anymore. Wayland runs fine on nvidia. You can get human readable compiler errors in C++. I wonder if part of the lack of excitement isn't how well a lot of things have started working and how mature software in general has become.
Also speaking personally, my own search engine has kinda been growing its userbase quite significantly this summer/fall. It's exciting for me at least :P
But I think in general you see what you're paying attention to, and being on the constant look-out for bad things is a bad mental habit. Thinking back 15 years ago, did we use to do this? I don't know if it's the loss of common cultural touchstones but it feels like there's so much tribalism everywhere now. The injection of clankers into every message board hasn't helped. But never has a place so populated felt so alienating as the web does in 2026.
I do really like Marginalia and I have been for a long time. But, to be honest, I’m not "excited" by it as an end user. Because it solves a problem that was already solved before but for which all existing solutions were enshitified.
So while I appreciate Marginalia a lot, while I believe it is very useful, while I find it awesome that a one-man shop is tackling what multinationals don’t really manage to do properly anymore, I’m not "excited" (but I fully understand that you are). Maybe it will come when I do realize that Marginalia bring something that doesn’t really exist yet: not the ability to search the web but the ability to get rid of slop when searching.
Wasn't that almost exactly what we were doing with Linux all along though, building a parallel ecosystem to the already existing and working Microsoft hegemony?
I guess the difference with internet search engines is you can't really host them yourself (without a prohibitive time and hardware investment), which takes away a lot of the tinkerability.
I think in many ways the big difference between then and now is a loss in sense of ownership. Tinkerability is at least some part in that, even if a lot of it comes from locked down hardware and DRM. Like even small things like being able to style the UI components in your applications is falling apart. You could do this 20 years ago, 15 years ago, 10 years ago, but now you at best get to choose between dark and light mode and it'll work on half your installed UI toolkits and some of the browsers you have installed.
This might come across as kind of weird, but whenever I feel the feelings you're describing of losing hope in tech and the world as a whole, I start browsing Neocities-type personal webpages. I follow every link. I swear, the passion some of these young people have, to continually put cool stuff on the web despite gestures vaguely, is kinda overwhelming; that usually fixes me. They're building webrings. They have personal Forgejos with cool projects. They share their Nix configs. They host friends-only fediverse instances. They unanimously hate AI with a passion. We're going to be okay.
When I look back at the things that excited me about tech, it often was the sheer "numbers go up" things. More hertz, more polygons. Despite Wirth's Law, this actually meant something. Code compiled faster, games got more beautiful, latency got down.
This stopped for me for several reasons. One the less rational end of things, the numbers aren't as tangible anymore. More cores instead of just more ghz, these day even "performance" cores. Some advances even peaked (phones can't really get smaller). Some advances just don't matter to me as much, as we've been in "good enough" phases for 10+ years. And quite often, Wirth's Law does apply quite well (compilation speed being a prime example).
I wonder what younger people see here, where they never had that "gigahertz hump" or maybe even smartphones with 128 megs of ram…
Things getting faster, also often was part in making things accessible to me. 3D stuff (rendering, modeling, gaming) no longer being constrained to expensive SGI workstations. Open Source was an even bigger contributor here: Starting with my first Linux CD-ROM I had access to so many programming languages and libraries, even back then. C, Forth, Ada, Perl, Python, all the Lisps!
These days? Well, I could probably spend a lot of money to do local LLMs, but otherwise?
Don't get me wrong, there's still new things for me to do, but it's rarely at the frontline of tech development (e.g. doing Arduino-style hardware hacking).
For me this also started long before the age of vibecoding. Probably when we entered the "cloud", and "architecture" was more about network diagrams than code design docs, Booch clouds or even UML. When desktop GUIs stopped being something one does. When open source projects suddenly meant JS libraries founded by tech giants (who thought I once miss the days when it was new editors/IRC clients).
Slop development doesn't make it easier, but it also doesn't feel new in that particular regard. The new ennui this creates is more on my "political annoyances" stack than on my "disappointed by tech" mountain.
So, I guess, faux-retrocomputing, where it's not about emulation but about "how we did things in the olden times of the 00s"?
I get excited by new Emacs releases, so much has been improved in my favorite editor. Although the third party stuff has a high slop-ratio, Emacs itself and the core stuff is human. And my favorite programming language Haskell also has had lots of QoL improvements lately, and again there are strong community values around understanding code (with exceptions of course, but I actually prefer a plurality of voices).
I used to find Machine Learning interesting. I no longer do. But I think my interest in it started waning some years before ChatGPT, when it seemed all the new methods required very expensive GPU setups and/or huge pre-trained models. I don't find it fun if I can't run it on my laptop.
Just want to say that I resonate with this a lot. Not a few years ago, I was always the one to explore shiny new things; now I dread all of these.
Nowadays, I follow a handful of indie game developers, graphics programmers, and tech artists working on small projects. Their enthusiasm sometimes lifts my spirits too.
Even if you are interested by AI, the ecological consequences and the way it is used are simply awful.
Yeah. I can relate to this one too. Machine learning was (and is still) fascinating to me, from an intellectual-curiosity perspective. Though my interest is definitely tainted by the current trend
Sorry to break it to you but if you were interested in technology for its own sake, then the path dependence to this point was baked in from the very beginning. It is a tool for control and extraction and it always has been.
I don't really see any strong arguments for your assertion...?
I can buy that totalitarianism and forced helplessness is a hard-to-avoid consequence of technological capability as it makes citizens so individually capable that eventually even a single individual could destroy civilizations. But that doesn't make technology a tool of control, but control a consequence of technology.
You're not missing something, it's mostly quite depressing.
There are some fun green shoots if you look at small things, mostly in the hobby and community-driven space. There's a lot of push-back at how rubbish things are getting and it's pushing everyday users toward libre and community-driven alternatives. Also, awareness of the dangers of tech (and in particular, its use as a force multiplier for the actions of awful people) is spreading more widely. The anti-AI pushback has gone global and it's something that I've found popping up in conversations with people I wouldn't previously have marked as 'techy'. People are looking for alternatives.
For me, the fun stuff is currently in the local-first and self-hosting movements, both of which have seen a lot of activity. It's surprisingly easy to run very mature tech stacks yourself, on your own infrastructure nowadays.
My 25 years in tech has gone from stressful, but loved, to 1000+ files per git pull. Its a complete AI clusterfk and taking any joy out of working in tech. Up top are tainted with the lies and promises from Anthropic and the likes. Quantity of code was never our issue, quality and maintainability was... and now we have buckets more code, and less maintainability and quality than ever. Metrics are tracked and we are always under the threat of binned if not 10x'ing ya shit with AI from c*.
I hate it all. Tech is destroyed, race to the bottom.
marginalia | 2 hours ago
Answering your question: Linux for gaming is kinda insane these days. With the exception of a small number of competitive multiplayer games where the devs are actively working to sabotage progress, pretty much anything runs as good or better on Linux now than it does on Windows. Even vintage games that Windows struggles to run.
Many other cases of things just improving over time to go from irredeemably broken to working so well you don't think about it anymore. Wayland runs fine on nvidia. You can get human readable compiler errors in C++. I wonder if part of the lack of excitement isn't how well a lot of things have started working and how mature software in general has become.
Also speaking personally, my own search engine has kinda been growing its userbase quite significantly this summer/fall. It's exciting for me at least :P
But I think in general you see what you're paying attention to, and being on the constant look-out for bad things is a bad mental habit. Thinking back 15 years ago, did we use to do this? I don't know if it's the loss of common cultural touchstones but it feels like there's so much tribalism everywhere now. The injection of clankers into every message board hasn't helped. But never has a place so populated felt so alienating as the web does in 2026.
[OP] ploum | an hour ago
Very wise words! Thanks for that.
I do really like Marginalia and I have been for a long time. But, to be honest, I’m not "excited" by it as an end user. Because it solves a problem that was already solved before but for which all existing solutions were enshitified.
So while I appreciate Marginalia a lot, while I believe it is very useful, while I find it awesome that a one-man shop is tackling what multinationals don’t really manage to do properly anymore, I’m not "excited" (but I fully understand that you are). Maybe it will come when I do realize that Marginalia bring something that doesn’t really exist yet: not the ability to search the web but the ability to get rid of slop when searching.
marginalia | an hour ago
Wasn't that almost exactly what we were doing with Linux all along though, building a parallel ecosystem to the already existing and working Microsoft hegemony?
I guess the difference with internet search engines is you can't really host them yourself (without a prohibitive time and hardware investment), which takes away a lot of the tinkerability.
I think in many ways the big difference between then and now is a loss in sense of ownership. Tinkerability is at least some part in that, even if a lot of it comes from locked down hardware and DRM. Like even small things like being able to style the UI components in your applications is falling apart. You could do this 20 years ago, 15 years ago, 10 years ago, but now you at best get to choose between dark and light mode and it'll work on half your installed UI toolkits and some of the browsers you have installed.
darkkindness | 49 minutes ago
This might come across as kind of weird, but whenever I feel the feelings you're describing of losing hope in tech and the world as a whole, I start browsing Neocities-type personal webpages. I follow every link. I swear, the passion some of these young people have, to continually put cool stuff on the web despite gestures vaguely, is kinda overwhelming; that usually fixes me. They're building webrings. They have personal Forgejos with cool projects. They share their Nix configs. They host friends-only fediverse instances. They unanimously hate AI with a passion. We're going to be okay.
mhd | an hour ago
When I look back at the things that excited me about tech, it often was the sheer "numbers go up" things. More hertz, more polygons. Despite Wirth's Law, this actually meant something. Code compiled faster, games got more beautiful, latency got down.
This stopped for me for several reasons. One the less rational end of things, the numbers aren't as tangible anymore. More cores instead of just more ghz, these day even "performance" cores. Some advances even peaked (phones can't really get smaller). Some advances just don't matter to me as much, as we've been in "good enough" phases for 10+ years. And quite often, Wirth's Law does apply quite well (compilation speed being a prime example).
I wonder what younger people see here, where they never had that "gigahertz hump" or maybe even smartphones with 128 megs of ram…
Things getting faster, also often was part in making things accessible to me. 3D stuff (rendering, modeling, gaming) no longer being constrained to expensive SGI workstations. Open Source was an even bigger contributor here: Starting with my first Linux CD-ROM I had access to so many programming languages and libraries, even back then. C, Forth, Ada, Perl, Python, all the Lisps! These days? Well, I could probably spend a lot of money to do local LLMs, but otherwise? Don't get me wrong, there's still new things for me to do, but it's rarely at the frontline of tech development (e.g. doing Arduino-style hardware hacking).
For me this also started long before the age of vibecoding. Probably when we entered the "cloud", and "architecture" was more about network diagrams than code design docs, Booch clouds or even UML. When desktop GUIs stopped being something one does. When open source projects suddenly meant JS libraries founded by tech giants (who thought I once miss the days when it was new editors/IRC clients).
Slop development doesn't make it easier, but it also doesn't feel new in that particular regard. The new ennui this creates is more on my "political annoyances" stack than on my "disappointed by tech" mountain.
So, I guess, faux-retrocomputing, where it's not about emulation but about "how we did things in the olden times of the 00s"?
Sanity | 52 minutes ago
I get excited by new Emacs releases, so much has been improved in my favorite editor. Although the third party stuff has a high slop-ratio, Emacs itself and the core stuff is human. And my favorite programming language Haskell also has had lots of QoL improvements lately, and again there are strong community values around understanding code (with exceptions of course, but I actually prefer a plurality of voices).
I used to find Machine Learning interesting. I no longer do. But I think my interest in it started waning some years before ChatGPT, when it seemed all the new methods required very expensive GPU setups and/or huge pre-trained models. I don't find it fun if I can't run it on my laptop.
LesleyLai | 56 minutes ago
Just want to say that I resonate with this a lot. Not a few years ago, I was always the one to explore shiny new things; now I dread all of these.
Nowadays, I follow a handful of indie game developers, graphics programmers, and tech artists working on small projects. Their enthusiasm sometimes lifts my spirits too.
Yeah. I can relate to this one too. Machine learning was (and is still) fascinating to me, from an intellectual-curiosity perspective. Though my interest is definitely tainted by the current trend
alper | 2 hours ago
Sorry to break it to you but if you were interested in technology for its own sake, then the path dependence to this point was baked in from the very beginning. It is a tool for control and extraction and it always has been.
The alternative is this approach: https://papanek.org/publications/designer-for-the-real/
marginalia | 2 hours ago
I don't really see any strong arguments for your assertion...?
I can buy that totalitarianism and forced helplessness is a hard-to-avoid consequence of technological capability as it makes citizens so individually capable that eventually even a single individual could destroy civilizations. But that doesn't make technology a tool of control, but control a consequence of technology.
zesterer | 56 minutes ago
You're not missing something, it's mostly quite depressing.
There are some fun green shoots if you look at small things, mostly in the hobby and community-driven space. There's a lot of push-back at how rubbish things are getting and it's pushing everyday users toward libre and community-driven alternatives. Also, awareness of the dangers of tech (and in particular, its use as a force multiplier for the actions of awful people) is spreading more widely. The anti-AI pushback has gone global and it's something that I've found popping up in conversations with people I wouldn't previously have marked as 'techy'. People are looking for alternatives.
For me, the fun stuff is currently in the local-first and self-hosting movements, both of which have seen a lot of activity. It's surprisingly easy to run very mature tech stacks yourself, on your own infrastructure nowadays.
greg_loscombe | 30 minutes ago
My 25 years in tech has gone from stressful, but loved, to 1000+ files per git pull. Its a complete AI clusterfk and taking any joy out of working in tech. Up top are tainted with the lies and promises from Anthropic and the likes. Quantity of code was never our issue, quality and maintainability was... and now we have buckets more code, and less maintainability and quality than ever. Metrics are tracked and we are always under the threat of binned if not 10x'ing ya shit with AI from c*.
I hate it all. Tech is destroyed, race to the bottom.