going all-in on Linux is one thing, but going all-in on a specific window manager? with specific keybinds? idk, individual workflows are too specific to be prescribed like this imo.
Actually Omarchy is also run by 37 signals so this post is effectively that them saying that they're going all in on the same stack as a group. They can obviously just change what that stack is whenever they want though. This is coming from the inventor of rails so it's not that surprising to see a push for conventions.
I had a similar thought, but at the same time, if people were mandated to use Windows or MacOS then that would also pretty much lock you into their respective window managers. I guess it feels more restrictive partly because it's more common to pick and choose WMs on linux. (And partly because, yeah, seems like the setup goes way beyond just a distro+WM).
Also because it's a niche WM catering to pretty specific preferences. I used tiled window managers for a while but eventually I decided I preferred the no-customization, one-size-fits-all Gnome/KDE experience. A hyperland config is going to fit like a tailor-made glove—but in this case, it's a glove tailor-made for your CEO, not you.
I spun up Omarchy in a VM just to see what the fuss was about, and when I opened Neovim it booted into a plugin manager and started installing at least two dozen random plugins, including an extremely over-eager autocomplete config that filled my screen with snippet suggestions. I was instantly irritated.
I've been running Omarchy since last week. The theming is fun.
Getting used to hyprland and walker has been a short and eye-opening experience. I've had tiling before on Windows, but I've never been forced to use the tiling exclusively.
I'm all in on the vision behind Omarchy. I recognize DHH has a bit of an enthusiastic bent - He's been obsessive about Omarchy for the past month, and his opinions change. In some ways, he lives by the mantra "strong opinions, weakly held". I don't think that's a flaw.
I get the sense that this one is "the one" -- It's a foundation of linux that is entirely dependent on opinions, and DHH has them.
But, I think more importantly, the groundswell around the project highlights that there's a general dissatisfaction with the state of operating systems. The move to Linux has long been overdue for me personally. I'm incredibly tired of Windows and the Microsoft shenanigans. The adware on what should be a personal computer is an abomination.
I see Omarchy as an opinionated way of composing Linux in such a way that it offers a uniquely different premise of what an OS can be.
So far, I'm loving it, but am still tethered to Windows for some work stuff.
On the note of all the linux marketing, Jonathan Blow summed it up best:
> The people who would historically be excited about a new operating system can't do that any more, because everyone is too helpless to even conceive of a new OS.
> So they have to get excited about a mildly different arrangement of bloatware from That OS From 35 Years Ago.
> But as long as you give it a Cool Name, everything is good.
> Elon: makes car company (when everyone thinks electric cars will never work), rocket company (the rockets land themselves), Neuromancer brain chip company.
> Computer Nerds: Noooooo I can’t make an OS because drivers and adoption!!!!1
———
And from another thread
> It would be nice to have an OS with a proper job system as a core component. No legacy threads or mutexes at all. Everything is designed to be fine-grained parallel for modern 16+ core CPUs.
> For starters, every API is asynchronous command buffers with an optional slower/easier noob API on top. There are a lot of things that could tremendously simplify userspace as well.
I saw J. Blow's post. I'd love to see a new OS. It's easy to get on the internet and provide armchair criticism, especially when you're as storied as he is, but how does that get funded? Who actually makes it happen? It is a fun thought experiment, but I fear for a world where some VC-backed startup creates a new pit to fall in.
If I've learned any lessons in building software, it's that you very often have to strangle the old to make way for the new, but you get stuck in hell if you try to do it all at once. Getting Linux adoption up and general support improved, and then leveraging it as a WSL-esque bootstrap into a new OS might be a way.
On Komorebi, put simply, as an OSS maintainer I think me and that maintainer see the world a very different way. Would use it if not for the approach to licensing - does look cool!
The Register called this in 2007 [1] when they claimed that the Psion 5MX was the last computer to be fully developed by one team (hardware, software and design).
Okay, but _you literally can't do it_. Google couldn't do it (Fuchsia is dead). Blow can't do it. Ripping the bottom layer out of the stack the world is built on is just too hard.
Someone is going to make a new OS based around LLMs soon. It might end up being mobile but desktop could work. Voice and tool calling are close to being good enough.
Probably the entire reason for his success tbh. Unbridled enthusiasm without the guardrails holding himself back combined with a bit of luck and its almost the perfect combo for entrepreneurism?
> I get about 6 hours of mixed use from my Framework 13, so whenever I suspect that might be a problem, I bring a small 20K mAh Anker battery in the bag, and now I have double the capacity.
That power bank is 1.2lbs, and the laptop is 2.9lbs. Carrying around 4.1lbs ain’t light.
Maybe I’m just getting old but I’ve long since moved past thinking it’s ok to be inconvenienced due to what I’d consider a product deficiency.
Note: I’m not hating on Framework. I’m assuming this is an inconvenience due to running Linux on the desktop.
Purely a Framework issue. There are plenty of Intel Lunar Lake laptops that are ~20hr battery nowadays and competitive with Apple and run Linux quite fine. Framework won't incorporate due to the soldered RAM and AMD isn't power competitive with that yet.
Bad battery life is not an inherent property of Linux. If it were, you wouldn't see so many "smart devices" running flavors of Linux.
With respect to laptops, there aren't enough people looking at every SKU, of every manufacturer, and then translating the possible power state combinations of each bus and link and adapter into the config tools preferred by this-or-that single distro. It would require be a massive, more-thankless-than-average undertaking. As a result, the median laptop is not optimized for battery life out of the box, even for distros that do try (to say nothing for those like Arch that intentionally eschew making choices for you).
With some tools that allow self-tuning, it's not difficult to get 8-10 hours of battery life on the newest x86 platforms (or more on ARM laptops). However, it isn't done for you. It is an area that desktop Linux needs to invest in improving on behalf of the majority of people, who are less inclined to experiment and tweak.
Probably a Framework thing, maybe Hyprland doesn't have great power management, dunno. He also got the beefiest APU Framework offers in a laptop so maybe that's it.
I have an MSI laptop with an 11th gen Intel chipset and 16inch screen. I run Debian Gnome and get an easy 6 hours of work and 4-5 hours of straight Netflix and nothing about the laptop screams power efficient...
Actually hilarious that he couldn't stand Windows for more than a few months. Omarchy looks cool but for me, Gnome has enough tiling features and is easy for my family to use. If I want to rice I rice my terminal and editor, which I know no one else will ever touch. Also not a fan of Arch, much prefer Debian.
1/3 of basecamp employees, at the peak of the tech hiring bubble, took extremely generous buyouts, when they werent happy that their work wasnt going to the political ngo they wished it was.
I think a big part of this is a reflection of DHH's (understandable) frustration with Apple's App Store management, processes, and policies.
Looking back at January 2024, it seems like DHH started to reach his personal boiling point with that sort of nonsense, which seems to mark the start of the aggressive search for another suitable platform, out from under Apple's control.
While the original pursuit may have been less about seeking Windows or Linux than it was about getting away from Apple, it does seem like the end destination in Linux has been pretty satisfying. And it's great to see someone investing time and energy into promoting a cohesive linux desktop user experience that isn't just focused on feature parity to something else.
You miss a lot just going by the headline. He makes it blindingly clear in the first paragraph that Windows wouldn't even be an option if not for the Linux subsystem.
I started seeing a lot of dhh talking about Omarchy on twitter. It's been a joy to see fellow hyprland user, vocal about hyprland and linux. Linux users have been talking that every year that it's going to be the year of linux desktop, but now it feels so close when a person with big influence talks loud on social media.
I checked my git-log from my dotfiles. It says I installed hyprland 2023-07-23.
I used to run i3, dwm, bspwm (my favorite on X11) and tinkered with other wms. Since this commit, I have been full-time on Wayland. When Hyprland were young, it certainly had rough edges. But these days I don't feel overwhelmed even when I update several releases at once. It is stable and just works. The creator of hyprland, vaxry, is an incredible developer and maintainer. He's made so much progress on usability of Wayland.
This year has been wild for "person with big influence talks linux" events. The major difference this year compared with previous is they've all had really positive experiences.
I love dhh and I applaud his manic obsession with Linux over the 18 months or so that I've seen on my twitter feed. I still don't exactly understand what the purpose of all of this is.
I'm sure it is very configurable, but every visual I've seen of this thing looks awful and not something I'd want to look at while working. But I understand we all have different tastes.
But even in the blog post I'm struggling with 'why?' here. Am I to understand the primary benefits here are improved battery life and increased developer productivity by tests running faster? Is that it?
I travel an inordinate amount and have never found a Macbook's battery life to be insufficient. I struggle to even remember the last time I've used my computer long enough to drain the batter and not be near a power outlet. I work from ski lodges, planes, my car. This has never been a problem for me. Not once. This just feels like a really bad metric to optimize for given a typical developers' schedule and work arrangement.
> On the flip side, we'll get a massive boost in productivity from being able to run our Ruby on Rails test suites locally much faster.
Is this not just a Ruby issue? I don't know what's basecamp or HEYs codebase looks like on the inside, but they don't feel like projects whose tests suites should require a completely different OS or hardware arrangement. I haven't used Ruby in a decade but I do recall it being frustratingly slow. This seemed to be an understood and accepted reality amongst teams that adopt it.
Anyway, I feel like a better 'why you should do this' in order, especially if it is being mandated amongst developers in a company.
Linux is fundamentally appealing to those who want to tinker with OS. Even with this kind of full-featured ricing or package, users will have hard time to figures some features out, if they do not have enough knowledge how to navigate Linux system.
I personally use all three major OS, Mac at work, Windows for gaming, Linux for everything else. For me different OS serves different purposes and gives me task-oriented productivity boost.
> Linux is fundamentally appealing to those who want to tinker with OS
I respect that and truly admire what dhh has done and tweeted about over the last year or so. But for a business to mandate an operating system for their developers (which I fully agree they have the right to do) with those reasons is underwhelming.
What is a productivity boost for you in one OS might be a burden to someone else. And vice versa. That's the benefit of letting developers choose their OS.
No one’s forcing anyone to work at this company. Thanks to the publicity there’s plenty of warning about what it’ll involve so you can expect folks to come in well informed - or self-select out.
>I haven't used Ruby in a decade but I do recall it being frustratingly slow.
I've been using for more than a decade and have never found it to be "frustratingly slow", whatever that means in practice. Except when running on Docker on a co-worker's Apple product, which I do not use.
I haven’t used Ruby much, but anecdotally I once rewrote a Ruby service in Java and was able to reduce the number of servers needed from ~100 to 10. That was a case where it was “frustratingly slow” in practice.
Anyway, obviously Ruby (or any other dynamic language for that matter) isn't the solution to every problem. I was once tasked to refurbish an old Delphi desktop application that did some thermal simulations; I wrote the interface as a Rails webapp and the simulation itself as a compiled C routine, which I called from Rails via FFI.
From what I understood, the initial motivation wasn't technical, but related to Apple's practices around a closed app store and the 30% tax on every purchase. He explained it both on Lex Fridman's podcast at [1] and a bit on [2].
DHH is a weirdo, obviously, but I'll say that if you're doing Unixy dev work, there's less pain doing it on Linux than on MacOS, in the same way that MacOS is better than WSL.
As a non-obsessed Linux user I can scarcely notice the difference between different distros. Every place I've worked needed macs because we had to target iOS for at least some percentage of products.
This has been one of the most surprising developments of late, given DHH's and Rails devs historical preferences for Apple products. I'd love to see some stats on the impact of this change.
After many years using Ubuntu, I migrated to Omarchy this weekend (Arch Linux + Hyperlnd, a tiling window-manager). Looking great so far!
I don't think it's a DHH thing. I've been all on Apple for almost 20 years, but plan to make the leap to Linux in the next 6 to 9 months. Apple doesn't feel like the same company they used to be. Is that just perception? I don't know, but the fact that I can't delete the News or Messages apps is ridiculous.
I think a lot of people are sick of supporting poopy companies.
What annoys me most is how they keep hiding the address bar everywhere so I don't know the location of my files anymore. As a developer, this drives me insane.
you certainly can. Omarchy doesn't appear to be anything special, just a tiling WM (of which there are plenty of on macOS: aerospace, yabai, amethyst, etc.) with some preinstalled applications + basic dotfiles. people have been running similar setups for years on macOS
Are you aware that Omarchy is just a bunch of scripts on top of a GNU/Linux distro, which has been available in one way or another for the past three decades?
DHH and 37Signals fall into the same traps as so many libertarians: in trying to avoid control of others, they exert undue control over people themselves (we won't be beholden to Apple, because our employees must use Linux), in trying to be in control of their own destiny they end up with home-grown but worse options unless the only goal is control (not using the cloud or cloud tooling and instead writing their own). In trying to remove politics from their workplace they end up making a very strong political statement in support of exclusion.
I'm not anti-libertarian, I share some of the values, and I don't even disagree with some of the technical choices at 37Signals. However I think there's a lot of hypocrisy in the space, and DHH frequently comes across as being very tone-deaf and unaware of the impact his decisions have on others, basically an extreme lack of empathy.
He is slowly devolving into something.... He also now has a vocal bend on parenting and all men should be fathers and need to have more children. Can't wait until that becomes a requirement to work at his company...
> in trying to be in control of their own destiny they end up with home-grown but worse options
This very well describes many things I saw working in big tech however. Perhaps fully owning an internal implementation adds more weight to the decision than purely quality of the product itself.
Most of these criticisms don't stand up to scrutiny, but I can understand how you'd arrive at this opinion based on the blog post alone.
In reality, this change primarily applies to the engineering team and will be phased in over several years. The company as a whole will continue to run Windows, macOS and Linux.
There's also a get-out clause for any developers who just can't get along with Linux and prefer to stick with their beloved Mac.
This seems wildly subjective. Can you give some examples of how Omarchy is worse?
I find using macOS to be generally frustrating and cumbersome to use and I'm regularly confused when people say it's unequivocally the better choice for all things.
My examples for the home-grown tools were things like their deployment tooling which is far less well developed than other tools in the ecosystem. However I can see Omarchy going the same way.
It's very easy to quickly build something that solves 80% of your problem, and because it's tailored for what you need and you understand it, it's easy to believe that's better than the existing tools. The problem comes in the long term, in supporting more use-cases. You can decide not to support more, but that's a sacrifice (and in the case of Omarchy the end users may bear the brunt of that), or you need to put the time in to develop it.
Omarchy will have less development effort than macOS, their deployment system will have less development than Docker/Terraform/etc, and that's all a trade-off.
> Omarchy will have less development effort than macOS, their deployment system will have less development than Docker/Terraform/etc, and that's all a trade-off.
But… their deployement IS docker and terraform and etc. Their platform is Linux, nvim, etc. In total there is MORE development than all of apples platforms. It’s just not unified under one umbrella.
Anyways you have a false equivalency there. Something being "less developed" doesn't automatically make it "worse". On the contrary, projects with larger active development effort generally trend to enshittification whereas smaller projects can stay focused on solving their core use cases.
Good point about no one batting an eyelid over Mac or Windows laptop mandate but people are shocked at a Linux mandate... But imagine their second shock when they realize how good Linux is nowadays. I always had the sense that Linux has been the victim of a conspiracy between other OS makers since its inception. It's been suppressed within enterprise environments and rarely discussed anywhere. Linux's success on the server side has been entirely organic, working against all media narratives and vested interests.
As a developer, I am much more productive with Linux than any other OS. The workspaces feature is better (on most distros I tried), the commands more closely align with the prod environment, no need for complex virtual machine setups, no weird delays or missing critical commands when using bash. Everything just works and performs.
IMO, this is the payoff of having architected the system properly from the ground up, it's more stable. You don't get so many weird glitches as you do on Windows.
Windows NT's architecture is perfectly fine and in some respects superior to Unix, but Microsoft has made subsequent compromises that reduced stability.
going all-in on Linux is one thing, but going all-in on a specific window manager? with specific keybinds? idk, individual workflows are too specific to be prescribed like this imo.
Actually Omarchy is also run by 37 signals so this post is effectively that them saying that they're going all in on the same stack as a group. They can obviously just change what that stack is whenever they want though. This is coming from the inventor of rails so it's not that surprising to see a push for conventions.
I had a similar thought, but at the same time, if people were mandated to use Windows or MacOS then that would also pretty much lock you into their respective window managers. I guess it feels more restrictive partly because it's more common to pick and choose WMs on linux. (And partly because, yeah, seems like the setup goes way beyond just a distro+WM).
Also because it's a niche WM catering to pretty specific preferences. I used tiled window managers for a while but eventually I decided I preferred the no-customization, one-size-fits-all Gnome/KDE experience. A hyperland config is going to fit like a tailor-made glove—but in this case, it's a glove tailor-made for your CEO, not you.
I spun up Omarchy in a VM just to see what the fuss was about, and when I opened Neovim it booted into a plugin manager and started installing at least two dozen random plugins, including an extremely over-eager autocomplete config that filled my screen with snippet suggestions. I was instantly irritated.
it's just an example of providing a hyper tuned tool for a given usecase. In this case, for hacking on the 37signals codebase.
What's so special about the 37signals codebase that you'd have trouble being productive on it running Gnome or KDE?
I bet people install it and then are like "okay, but I tweaked it up to use GNOME/Plasma/whatever"
i suspect this is exactly what will happen. most people don't want hyprland lol.
I've been running Omarchy since last week. The theming is fun.
Getting used to hyprland and walker has been a short and eye-opening experience. I've had tiling before on Windows, but I've never been forced to use the tiling exclusively.
I'm all in on the vision behind Omarchy. I recognize DHH has a bit of an enthusiastic bent - He's been obsessive about Omarchy for the past month, and his opinions change. In some ways, he lives by the mantra "strong opinions, weakly held". I don't think that's a flaw.
I get the sense that this one is "the one" -- It's a foundation of linux that is entirely dependent on opinions, and DHH has them.
But, I think more importantly, the groundswell around the project highlights that there's a general dissatisfaction with the state of operating systems. The move to Linux has long been overdue for me personally. I'm incredibly tired of Windows and the Microsoft shenanigans. The adware on what should be a personal computer is an abomination.
I see Omarchy as an opinionated way of composing Linux in such a way that it offers a uniquely different premise of what an OS can be.
So far, I'm loving it, but am still tethered to Windows for some work stuff.
Try komorebic on windows — great tiling.
On the note of all the linux marketing, Jonathan Blow summed it up best:
> The people who would historically be excited about a new operating system can't do that any more, because everyone is too helpless to even conceive of a new OS.
> So they have to get excited about a mildly different arrangement of bloatware from That OS From 35 Years Ago.
> But as long as you give it a Cool Name, everything is good.
> Elon: makes car company (when everyone thinks electric cars will never work), rocket company (the rockets land themselves), Neuromancer brain chip company.
> Computer Nerds: Noooooo I can’t make an OS because drivers and adoption!!!!1
———
And from another thread
> It would be nice to have an OS with a proper job system as a core component. No legacy threads or mutexes at all. Everything is designed to be fine-grained parallel for modern 16+ core CPUs.
> For starters, every API is asynchronous command buffers with an optional slower/easier noob API on top. There are a lot of things that could tremendously simplify userspace as well.
I saw J. Blow's post. I'd love to see a new OS. It's easy to get on the internet and provide armchair criticism, especially when you're as storied as he is, but how does that get funded? Who actually makes it happen? It is a fun thought experiment, but I fear for a world where some VC-backed startup creates a new pit to fall in.
If I've learned any lessons in building software, it's that you very often have to strangle the old to make way for the new, but you get stuck in hell if you try to do it all at once. Getting Linux adoption up and general support improved, and then leveraging it as a WSL-esque bootstrap into a new OS might be a way.
On Komorebi, put simply, as an OSS maintainer I think me and that maintainer see the world a very different way. Would use it if not for the approach to licensing - does look cool!
The Register called this in 2007 [1] when they claimed that the Psion 5MX was the last computer to be fully developed by one team (hardware, software and design).
[1]: https://www.theregister.com/2007/06/26/psion_special/
Okay, but _you literally can't do it_. Google couldn't do it (Fuchsia is dead). Blow can't do it. Ripping the bottom layer out of the stack the world is built on is just too hard.
Someone is going to make a new OS based around LLMs soon. It might end up being mobile but desktop could work. Voice and tool calling are close to being good enough.
His ego never ceases to amaze me. He’s quite convinced everything he does is important and other people can’t wait to understand.
Probably the entire reason for his success tbh. Unbridled enthusiasm without the guardrails holding himself back combined with a bit of luck and its almost the perfect combo for entrepreneurism?
What lead you to this conclusion?
> I get about 6 hours of mixed use from my Framework 13, so whenever I suspect that might be a problem, I bring a small 20K mAh Anker battery in the bag, and now I have double the capacity.
That power bank is 1.2lbs, and the laptop is 2.9lbs. Carrying around 4.1lbs ain’t light.
Maybe I’m just getting old but I’ve long since moved past thinking it’s ok to be inconvenienced due to what I’d consider a product deficiency.
Note: I’m not hating on Framework. I’m assuming this is an inconvenience due to running Linux on the desktop.
Purely a Framework issue. There are plenty of Intel Lunar Lake laptops that are ~20hr battery nowadays and competitive with Apple and run Linux quite fine. Framework won't incorporate due to the soldered RAM and AMD isn't power competitive with that yet.
Bad battery life is not an inherent property of Linux. If it were, you wouldn't see so many "smart devices" running flavors of Linux.
With respect to laptops, there aren't enough people looking at every SKU, of every manufacturer, and then translating the possible power state combinations of each bus and link and adapter into the config tools preferred by this-or-that single distro. It would require be a massive, more-thankless-than-average undertaking. As a result, the median laptop is not optimized for battery life out of the box, even for distros that do try (to say nothing for those like Arch that intentionally eschew making choices for you).
With some tools that allow self-tuning, it's not difficult to get 8-10 hours of battery life on the newest x86 platforms (or more on ARM laptops). However, it isn't done for you. It is an area that desktop Linux needs to invest in improving on behalf of the majority of people, who are less inclined to experiment and tweak.
Probably a Framework thing, maybe Hyprland doesn't have great power management, dunno. He also got the beefiest APU Framework offers in a laptop so maybe that's it.
I have an MSI laptop with an 11th gen Intel chipset and 16inch screen. I run Debian Gnome and get an easy 6 hours of work and 4-5 hours of straight Netflix and nothing about the laptop screams power efficient...
Previously:
> March 7, 2024: Committing to Windows
https://world.hey.com/dhh/committing-to-windows-2d6388fd
> June 6, 2024: Introducing Omakub (based on Ubuntu)
https://world.hey.com/dhh/introducing-omakub-354db366
Place your bets on what is next: nixOS? Haiku? OpenBSD?
Losing patience with windows after a month or two sounds about right. Customised arch is probably a local optimum.
I haven't seen his earlier post about Windows. Kinda funny. He's put quite a lot of effort into Omarchy now though.
Actually hilarious that he couldn't stand Windows for more than a few months. Omarchy looks cool but for me, Gnome has enough tiling features and is easy for my family to use. If I want to rice I rice my terminal and editor, which I know no one else will ever touch. Also not a fan of Arch, much prefer Debian.
Previously:
> “About one-third of Basecamp employees accepted buyouts today” https://news.ycombinator.com/item?id=26998127
What's the relevance of this? Also, in many podcasts since, DHH and Jason Fried said this is one of the best things that they ever did.
1/3 of basecamp employees, at the peak of the tech hiring bubble, took extremely generous buyouts, when they werent happy that their work wasnt going to the political ngo they wished it was.
I think a big part of this is a reflection of DHH's (understandable) frustration with Apple's App Store management, processes, and policies.
Looking back at January 2024, it seems like DHH started to reach his personal boiling point with that sort of nonsense, which seems to mark the start of the aggressive search for another suitable platform, out from under Apple's control.
https://world.hey.com/dhh/apple-rejects-the-hey-calendar-fro...
https://world.hey.com/dhh/apple-s-new-extortion-regime-to-ke...
While the original pursuit may have been less about seeking Windows or Linux than it was about getting away from Apple, it does seem like the end destination in Linux has been pretty satisfying. And it's great to see someone investing time and energy into promoting a cohesive linux desktop user experience that isn't just focused on feature parity to something else.
You miss a lot just going by the headline. He makes it blindingly clear in the first paragraph that Windows wouldn't even be an option if not for the Linux subsystem.
I started seeing a lot of dhh talking about Omarchy on twitter. It's been a joy to see fellow hyprland user, vocal about hyprland and linux. Linux users have been talking that every year that it's going to be the year of linux desktop, but now it feels so close when a person with big influence talks loud on social media.
I checked my git-log from my dotfiles. It says I installed hyprland 2023-07-23.
> 319649c 2023-07-23 (sbinnee) install(hyprland): wayland wm tag: hyprland
I used to run i3, dwm, bspwm (my favorite on X11) and tinkered with other wms. Since this commit, I have been full-time on Wayland. When Hyprland were young, it certainly had rough edges. But these days I don't feel overwhelmed even when I update several releases at once. It is stable and just works. The creator of hyprland, vaxry, is an incredible developer and maintainer. He's made so much progress on usability of Wayland.
This year has been wild for "person with big influence talks linux" events. The major difference this year compared with previous is they've all had really positive experiences.
I love dhh and I applaud his manic obsession with Linux over the 18 months or so that I've seen on my twitter feed. I still don't exactly understand what the purpose of all of this is.
I'm sure it is very configurable, but every visual I've seen of this thing looks awful and not something I'd want to look at while working. But I understand we all have different tastes.
But even in the blog post I'm struggling with 'why?' here. Am I to understand the primary benefits here are improved battery life and increased developer productivity by tests running faster? Is that it?
I travel an inordinate amount and have never found a Macbook's battery life to be insufficient. I struggle to even remember the last time I've used my computer long enough to drain the batter and not be near a power outlet. I work from ski lodges, planes, my car. This has never been a problem for me. Not once. This just feels like a really bad metric to optimize for given a typical developers' schedule and work arrangement.
> On the flip side, we'll get a massive boost in productivity from being able to run our Ruby on Rails test suites locally much faster.
Is this not just a Ruby issue? I don't know what's basecamp or HEYs codebase looks like on the inside, but they don't feel like projects whose tests suites should require a completely different OS or hardware arrangement. I haven't used Ruby in a decade but I do recall it being frustratingly slow. This seemed to be an understood and accepted reality amongst teams that adopt it.
Anyway, I feel like a better 'why you should do this' in order, especially if it is being mandated amongst developers in a company.
Linux is fundamentally appealing to those who want to tinker with OS. Even with this kind of full-featured ricing or package, users will have hard time to figures some features out, if they do not have enough knowledge how to navigate Linux system.
I personally use all three major OS, Mac at work, Windows for gaming, Linux for everything else. For me different OS serves different purposes and gives me task-oriented productivity boost.
> Linux is fundamentally appealing to those who want to tinker with OS
I respect that and truly admire what dhh has done and tweeted about over the last year or so. But for a business to mandate an operating system for their developers (which I fully agree they have the right to do) with those reasons is underwhelming.
What is a productivity boost for you in one OS might be a burden to someone else. And vice versa. That's the benefit of letting developers choose their OS.
No one’s forcing anyone to work at this company. Thanks to the publicity there’s plenty of warning about what it’ll involve so you can expect folks to come in well informed - or self-select out.
Strawman. Nobody ever said anyone was being forced to work there and nobody would claim that was the case.
>I haven't used Ruby in a decade but I do recall it being frustratingly slow.
I've been using for more than a decade and have never found it to be "frustratingly slow", whatever that means in practice. Except when running on Docker on a co-worker's Apple product, which I do not use.
I haven’t used Ruby much, but anecdotally I once rewrote a Ruby service in Java and was able to reduce the number of servers needed from ~100 to 10. That was a case where it was “frustratingly slow” in practice.
Wasn't JRuby an option?
Anyway, obviously Ruby (or any other dynamic language for that matter) isn't the solution to every problem. I was once tasked to refurbish an old Delphi desktop application that did some thermal simulations; I wrote the interface as a Rails webapp and the simulation itself as a compiled C routine, which I called from Rails via FFI.
What is that service doing? Even 10 Java servers seems like a lot for a single service, unless 1 server = 1 single VM thread or something...
> Is this not just a Ruby issue?
Pretty sure it's a Docker on Mac issue.
From what I understood, the initial motivation wasn't technical, but related to Apple's practices around a closed app store and the 30% tax on every purchase. He explained it both on Lex Fridman's podcast at [1] and a bit on [2].
[1] https://www.youtube.com/watch?v=mzDi8u3WMj0 [2] https://world.hey.com/dhh/living-with-linux-and-android-afte...
Does that affect developer productivity at 37signals though?
It does, by making waves about Linux and making Linux better over time.
What is the “this” when you say “why you should do this”?
DHH is a weirdo, obviously, but I'll say that if you're doing Unixy dev work, there's less pain doing it on Linux than on MacOS, in the same way that MacOS is better than WSL.
As a non-obsessed Linux user I can scarcely notice the difference between different distros. Every place I've worked needed macs because we had to target iOS for at least some percentage of products.
This has been one of the most surprising developments of late, given DHH's and Rails devs historical preferences for Apple products. I'd love to see some stats on the impact of this change.
After many years using Ubuntu, I migrated to Omarchy this weekend (Arch Linux + Hyperlnd, a tiling window-manager). Looking great so far!
I don't think it's a DHH thing. I've been all on Apple for almost 20 years, but plan to make the leap to Linux in the next 6 to 9 months. Apple doesn't feel like the same company they used to be. Is that just perception? I don't know, but the fact that I can't delete the News or Messages apps is ridiculous.
I think a lot of people are sick of supporting poopy companies.
Same here. Also moving from iOS to GrapheneOS (writing this on a GrapheneOS Pixel Tab instead of my usual iPad Pro).
After 6 years using Linux I've made my way back to Mac. Mac just has better software, things are more seamless
[dead]
Likewise. Used Macs for 20 years, and then switched to Linux (and the Framework laptop) about a year ago. Zero regrets.
What annoys me most is how they keep hiding the address bar everywhere so I don't know the location of my files anymore. As a developer, this drives me insane.
Use whatever dhh recommended last year!
I like parts of omakub, but impossible to use alacritty on Windows. Paste is broken. Known WONTFIX.
I would use a tiling window manager, but not ready to give up on the mouse. Title bars are sometimes nice.
Lazyvim not for experienced vi users.
The timing on this sucks. I would love to see what DHH would do with the new https://github.com/apple/container thing coming in OSX26
You can't do anything like Omarchy with macOS. Everything is locked down tight, no customization or tweaking allowed.
Lots of developers are tired of being hemmed in and disrespected by Apple. Omarchy gets us back into using an OS made for developers, by developers.
> You can't do anything like Omarchy with macOS
you certainly can. Omarchy doesn't appear to be anything special, just a tiling WM (of which there are plenty of on macOS: aerospace, yabai, amethyst, etc.) with some preinstalled applications + basic dotfiles. people have been running similar setups for years on macOS
Are you aware that Omarchy is just a bunch of scripts on top of a GNU/Linux distro, which has been available in one way or another for the past three decades?
DHH and 37Signals fall into the same traps as so many libertarians: in trying to avoid control of others, they exert undue control over people themselves (we won't be beholden to Apple, because our employees must use Linux), in trying to be in control of their own destiny they end up with home-grown but worse options unless the only goal is control (not using the cloud or cloud tooling and instead writing their own). In trying to remove politics from their workplace they end up making a very strong political statement in support of exclusion.
I'm not anti-libertarian, I share some of the values, and I don't even disagree with some of the technical choices at 37Signals. However I think there's a lot of hypocrisy in the space, and DHH frequently comes across as being very tone-deaf and unaware of the impact his decisions have on others, basically an extreme lack of empathy.
He is slowly devolving into something.... He also now has a vocal bend on parenting and all men should be fathers and need to have more children. Can't wait until that becomes a requirement to work at his company...
> in trying to be in control of their own destiny they end up with home-grown but worse options
This very well describes many things I saw working in big tech however. Perhaps fully owning an internal implementation adds more weight to the decision than purely quality of the product itself.
Most of these criticisms don't stand up to scrutiny, but I can understand how you'd arrive at this opinion based on the blog post alone.
In reality, this change primarily applies to the engineering team and will be phased in over several years. The company as a whole will continue to run Windows, macOS and Linux.
There's also a get-out clause for any developers who just can't get along with Linux and prefer to stick with their beloved Mac.
Source: the podcast episode linked from the DHH post (https://37signals.com/podcast/moving-to-omarchy/).
> they end up with home-grown but worse options
This seems wildly subjective. Can you give some examples of how Omarchy is worse?
I find using macOS to be generally frustrating and cumbersome to use and I'm regularly confused when people say it's unequivocally the better choice for all things.
My examples for the home-grown tools were things like their deployment tooling which is far less well developed than other tools in the ecosystem. However I can see Omarchy going the same way.
It's very easy to quickly build something that solves 80% of your problem, and because it's tailored for what you need and you understand it, it's easy to believe that's better than the existing tools. The problem comes in the long term, in supporting more use-cases. You can decide not to support more, but that's a sacrifice (and in the case of Omarchy the end users may bear the brunt of that), or you need to put the time in to develop it.
Omarchy will have less development effort than macOS, their deployment system will have less development than Docker/Terraform/etc, and that's all a trade-off.
> Omarchy will have less development effort than macOS, their deployment system will have less development than Docker/Terraform/etc, and that's all a trade-off.
But… their deployement IS docker and terraform and etc. Their platform is Linux, nvim, etc. In total there is MORE development than all of apples platforms. It’s just not unified under one umbrella.
I asked about "worse", not "home-grown".
Anyways you have a false equivalency there. Something being "less developed" doesn't automatically make it "worse". On the contrary, projects with larger active development effort generally trend to enshittification whereas smaller projects can stay focused on solving their core use cases.
For a macOS-specific example, see https://x.com/ryolu_/status/1954598993515962707
I’m pretty sure self hosting is really just a cost thing for them.
[dead]
Good point about no one batting an eyelid over Mac or Windows laptop mandate but people are shocked at a Linux mandate... But imagine their second shock when they realize how good Linux is nowadays. I always had the sense that Linux has been the victim of a conspiracy between other OS makers since its inception. It's been suppressed within enterprise environments and rarely discussed anywhere. Linux's success on the server side has been entirely organic, working against all media narratives and vested interests.
As a developer, I am much more productive with Linux than any other OS. The workspaces feature is better (on most distros I tried), the commands more closely align with the prod environment, no need for complex virtual machine setups, no weird delays or missing critical commands when using bash. Everything just works and performs.
IMO, this is the payoff of having architected the system properly from the ground up, it's more stable. You don't get so many weird glitches as you do on Windows.
Printing is also better.
Windows NT's architecture is perfectly fine and in some respects superior to Unix, but Microsoft has made subsequent compromises that reduced stability.
He's such a an attention-addicted dweeb. Like Elon Musk if he could code. Shock jock CEO. It's tiresome.
[dead]