A useful way to frame this isn’t “is it worth tens of hours to avoid a future reinstall” but “where do I want my entropy to live”. You’re going to invest time somewhere: either in a slowly-accumulating pile of invisible state (brew, manual configs, random installers) or in a config that you can diff, review and roll back. The former feels free until you hit some cursed PATH/SSL/toolchain issue at 11pm and realize you’ve been paying that tax all along, just in tiny, forgotten increments.
I think where Nix shines isn’t “one laptop every 6 years” but when your environment needs to be shared or recreated: multiple machines, a team, or a project with nasty native deps. At that point, nix-darwin + dev shells becomes infrastructure, not a hobby. You don’t have to go all-in on “my whole Mac is Nix” either: keep GUI apps and casual tools imperative, and treat Nix as the source of truth for the stuff that actually blocks you from doing work. That hybrid model matches what the article hints at and tends to give you most of the upside without turning your personal laptop into a second job.
One of the biggest annoyances I have with doing this with Nix vs another tool is that Nix doesn't natively communicate back state changes so that you can make them reproducible.
If I make a git repo, place '~/.config/newsapp.conf' in there and then symlink it back to '~/.config/', if NewsApp introduces a new variable in its settings I am immediately aware because Git will complain about being dirty. However, Nix will happily just nuke that .conf and rebuild whatever is in your configuration, without letting you know about state. Which is ultimately bad for reproducibility. It's a huge blind spot in Nix.
It sounds like you're talking about home-manager, which is a third-party Nix module, not Nix itself. I've been using Nix happily for several years now for work and personal without needing to use home-manager at all (since I don't like it either).
I tried getting Nix working a couple of months ago and ditched it because changing some Tower settings updates the global gitconfig (as it should in this circumstance) and Nix would wipe them out.
All of it seemed way too annoying compared to just having a dotfiles repo, and if it couldn’t handle the Tower/gitconfig issue I know for sure everything else I was going to run into wasn’t worth it.
>I think where Nix shines isn’t “one laptop every 6 years” but when your environment needs to be shared or recreated: multiple machines, a team, or a project with nasty native deps.
I'd like to add the third thing, which is just iteration. It's very tricky to maintain advanced workflows even locally. I'd guess many won't even try to compose things that could work in combination (often self-hosted services), when they know they can't reliably maintain those environments.
I've tried nix-darwin a time or two in the past. Every few years when homebrew makes a "hostile" change and I get upset I consider trying it again (now most recently with changes to gatekeeper). I think I'll get to doing so in the next year or so.
But I think just in fairness, the comparison here for flakes should be to Homebrew bundles. My packages are managed in a bundle: https://github.com/Julian/dotfiles/blob/main/Brewfile and then locked by a lockfile: https://github.com/Julian/dotfiles/blob/main/Brewfile.lock.j... and installing is just `brew bundle install`. All native Homebrew functionality. In practice I have never had an issue with non-reproducible builds across my machines (partly because the tendency on macOS is to run the latest versions of things and stay up to date).
(But again I do find nix-darwin interesting to try for other reasons.)
Very nice, I think I'll be moving my "must-have" homebrew packages to a Brewfile. FYI tho, Homebrew no longer supports Brewfile.lock.json (it was always just a log anyway, not a lockfile). https://github.com/Homebrew/homebrew-bundle/pull/1509
> the comparison here for flakes should be to Homebrew bundles.
The bundler integration for nix-darwin actually just bakes tightly-controlled Brewfiles. It’s still worthwhile though, since part of the “tightly-controlled” means better cleanup when you remove things.
> The consequence is me, spending a few hours debugging my environment instead of writing code.
But then I also see this:
> I’ve spent a lot of time recently moving my entire workflow into a declarative system using nix.
I can see how this can be beneficial for someone who switches systems very often, reinstalls their OS from scratch very often, or just derives a lot of pleasure/peace of mind knowing that their dev env is immutable.
I change computers once every 6 years or so, maybe more. To me this looks like exchanging a couple (hypothetical) hours of debugging 6 years in the future by tens of (guaranteed) hours trying to climb up the nix learning cliff.
I am happy that it works for the author though, and knowing that it's possible is good in case my particular development circumstances change.
Nix is not worth it if all you want is configuring your home computer. The learning curve is steep and has a tall onboarding cliff.
The only way you get positive ROI from Nix is either you enjoy the journey, or you use it to do more than just managing a single computer: you manage a fleet, you build thin application container images, you bundle all your software, you have devshells, repeatable tests and deploys, etc. It's the same tool for all of them.
It's worth it for me. I learnt and set up my home manager config once in 2024 and I now only occasionally make light tweaks. But I deploy it on every personal computer or remote development servers at my jobs (previous and current). Granted, I'm probably one of the handful of people in my community who uses nix, because it's too arcane - but I've already paid the cost.
Nix is a wonderful technology. But I would not argue it is practical if you can afford "just fix it when it breaks". A nix setup more/less requires you to pay all the cost up front.
I appreciate putting in the effort now so that I don't have to later for stuff like declarative dev environments. It's really nice to not have to copy-and-paste installation instructions from a README. -- I did like the point: until you've felt what a comfortable design is, you cannot imagine it.
Yeah, I'm a Nix user and fan but I'm still going to be stuck fixing it when it breaks. The biggest thing that I have to remind myself is that reproducible != "just works." Once you get to what you want, you're set but until then and every change after, there's a chance you'll be in the weeds.
At least from my usage, the fact that your configuration is all tied to whether the entirety of nixpkgs-unstable is working can be a real headache.
Like recently when CMake was upgraded to 4.0, it took down a healthy handful of packages, which meant you couldn't update anything until things were resolved or you were really fluent in Nix hackery. (I was the former)
Long-time (by now) Nix user here. It's even worse than a few hours of debugging. In my experience it requires continuous maintenance, options tend to be deprecated and moved around on a regular basis (at least in NixOS and home-manager), things like NeoVim break regularly as a result of friction between the immutable world and a lot of plugins expecting a mutable world, etc. I do install my Mac from scratch every 6 months or so (I like a clean system), but with a dotfiles repo and Homebrew, it takes me about an hour to set everything up again, which is far less time than maintaining a Nix-managed system.
For me Nix/NixOS is by far the most effective for deploying servers, development VMs, etc. I do this regularly, so it pays off - I have a familiar environment completely set up in minutes. Another place it pays off, even on macOS, is for development environments, especially if you have to pull in a lot of native dependencies (which can happen in mixed Python + C++ projects, mixed Rust + C/C++ projects, etc.).
Nix shines in all difficult cases, like setting up complex cross-compilation environments, building for old glibc versions, etc. You set it up once and then it's really easy for yourself and other project contributors to get the same build environment.
Like any tool, a good engineer knows when to apply it.
In the six years you are using your computer, do you ever expect to run into versioning issues and conflicts? Homebrew packages conflicting with local packages, something you compile give needs a different python/ruby/node/rust/whatever version that you have locally installed, you want to quickly try out a new package or upgrade without changing your system but have the option of rolling back safely, need to quickly install a database, want to try out a new shell and shell config but don't brick your system and have the option to roll back, etc. Nix gives you all of that and more for a one-time setup cost. Your argument is correct only if you expect to never change anything on your computer for the 6 years. But if I think about how often I have fought with homebrew or some kind of versioning/path/binary conflicts in the past then the investment in nix has paid off exponentially.
It's also about peace of mind like you said. Before nix I sometimes felt anxiety installing or upgrading certain things on my computer. "Will this upgrade break stuff?" - and often it did and I'd have to spend the next few hours debugging. With nix I don't worry about any of that anymore.
That's mostly solved with env managers for python/ruby/node/..., takes at most a few minutes to fully set up and learn, and doesn't get constantly broken by macOS updates.
Even for things like trying out a new shell you can temporarily move the dotfiles somewhere and restore them back and it still takes less time than converting everything to Nix.
But now you’re stuck with Python. Nix enables trivially simple dev environments that are completely heterogenous. This gives you a powerful form of freedom because it literally opens up the entire software universe to your dev environment in a confidence inspiring way. Not to mention things like parameterising anything you use reliably and setting up environment variables, shell scripts, database service whatever you want. Also integrates with tools such as UV really well. Yes, the language is terse and difficult but once you know it, it’s liberating, and makes you a better software developer in my opinion because you now have a high-end full workshop rather than a small toolbox.
Fully spot on, I don't get what is that hard to set a couple of environment variables, and mayby symbolic links, depending on the OS and language being used.
A simple UNIX script or PowerShell utility takes care of it.
None of the ones I have used during the last decades has ever grown to more than like 20 lines of code, minus comments.
This is my feeling too. Nix is a relatively high time investment for a tool that tries to do everything, when you might not need or want everything and using the specific language’s tooling is more than sufficient and quicker. It takes a few minutes to install and do `uv sync`, or `nvm install`, or whatever, on a repository on a new computer, and it just works. Until Nix gets there, and I’m skeptical it will because of the “purist” mindset a lot of people in the community have, it’s hard to justify it.
I think the comparison is "X-as-code", like with Terraform and other tools.
If you just want a throwaway VM, it's straightforward to create one through the UI cloud console. Whereas, terraform is nevertheless still a useful tool to use to manage VMs.
For stuff like installing development dependencies.. it's maybe not difficult to copy-and-paste instructions from a readme, but solutions like devcontainers or Nix's development shells can be useful even if costing more overhead.
until you need to start combining things. Docker is conceptually a VM the encapsulates everything nicely, but it ironically doesn't "compose" nearly as well as nix flakes or shells. With Nix you start out with a base env and can trivially extend it hierarchically and jump up and down the tree super easily, and without having to roll your own microservice architecture each time just to get stuff to work together.
Docker OTOH composes whole services nicely: if my project needs a redis cache and postgres instance, I don't have to faff about with local ports, traefik can pick up my web server, and so on. I use a flake to create and lock a local development toolchain, but it's no help in managing the services.
One thing I haven't tried yet is building a container from a flake, which would have obvious benefits for reproducibility. Still don't think it would help with service orchestration though.
> I change computers once every 6 years or so, maybe more. To me this looks like exchanging a couple (hypothetical) hours of debugging 6 years in the future by tens of (guaranteed) hours trying to climb up the nix learning cliff.
yes, it sounds like it's not worth it for you -- you will have to spend a significant amount of time converting your system to do things "the nix way". you can try to do this incrementally, but it's a time sink, and really easy to get stuck bikeshedding instead of doing work.
for me, it feels like a near equal trade-off between debugging nix, or debugging some random env issues that pop up. i know nix, claude code "knows" nix, a lot of other people online know nix. random env issues are random, and yield worse results on google, and frankly are much more frustrating to the point i would rather spend more time with nix than deal with them. maybe a very weird view.
I have experienced a positive return on investment from using Nix Darwin and devenv.sh since getting a new Mac two years ago. Did not spend too much time learning neither.
I do this too. It's not for everyone. At this point it's easily been positive ROI for me, but that's after about two years now of maintaining my configs through multiple machines and MacOS upgrades.
I would recommend it only if this type of thing naturally interests you. I can't imagine powering through the initial learning curve if it felt like a frustrating chore.
That said, if having (most of) your machine defined declaratively in a git repository sounds exciting/useful/comfy, then I would encourage you to give it a try. You can start small by just configuring a few programs or options and see how you like it.
only thing i'd add is mas for mac app store apps you want to ensure are installed but otherwise i run pretty much the same setup.
When i install a fresh macos i have two commands to run - install nix using the determinate systems installer, then apply my nix config.
It's not quite as streamlined as nixos but good enough.
My biggest remaining pain point is dev envs - i've been leaning into adding a flake in each project, so for example i have a single project that's written in scala 2.13, when i cd into that project dir, the correct jvm version, sbt, intellij etc are installed, some useful env vars and shell aliases etc. - that's all great (i haven't felt the need to adopt denenv.sh or flox yet) but i do find myself wanting a devcontainer sandbox workflow more often these days (blame cli coding "agents"), i lean on vscode for that rather than nix so far. In python (where i spend a lot more time) uv loses a lot of value in nix and i don't like that.
The real game changer after I went full nix OS on all my machines and using it on a a few Mac’s is indeed devshells which seemed so annoying to learn but really really are a game changer
I did this for a while but MacOS updates broke Nix often enough that I usually would spent some time every week reinstalling it. I still use Nix for dev environments because it is great but Nix still breaks sometimes.
I also really wanted to like the declarative homebrew configuration but it also often didn’t work as expected for some configurations and had a lot of leaky abstractions that straight up just broke sometimes.
If I ever go back to managing my Mac with nix I would probably just do a home-manager setup and just install most of the applications imperatively.
Given this was using an intel based machine around the time when the switch to arm came so a lot of breakage also stemmed from that.
I still use nix to handle my homelab.
My setup up on my Mac is as follows:
- Orbstack
- NixOS machine run in orbstack
- My whole dev environment is run from this container and is very transportable
- GUI apps are installed on my Mac using the App Store or homebrew etc. but I try to reduce the amount of installed applications
- if I have to install something that I don’t want to install but have to, I try to do it in a UTM machine.
Ive been using nix-darwin for over a year now after using nixos with flakes for a bit. I now have a singular repo with multiple machine configurations. Nixos for my home server, nix darwin on a macbook air and a nix darwin with a work config. This allows me to have common programs on all machines but also overlay some specialised packages and programs in certain environments. After climbing the initial mountain, its been very satisfying and things just work. My work laptop died recently and I was able to be fully up to speed in a fraction of the time it would have taken me otherwise.
It would be nice to set up immutable config down to the granularity of a git commit, sharing it as a single file, and be able to reproduce that setup on anyone's machine. Instant onboarding and mirrored production (barring secrets).
Recent developments have seen the creation of bootc images for non-Fedora distros too, and at this point I've seen quite a few cool arch-bootc custom images, completely customized to the author's desires. See: https://github.com/bootcrew/, https://github.com/tartaria-dev/tartaria
For me, this is the holy grail. Every time I switch laptops, I lift all my config files and such over, but there's always so many system level configs and other things that you have to go and manually fiddle. On top of that, some apps don't really behave well when you just move config files to e.g. under a different username etc. Would be nice if there was a comprehensive solution to this problem, need to try nix-darwin out.
nix-darwin is essentially this. I have a small bootstrap script to install Xcode CLI and Nix, git clone my dotfiles and activate the config. That in turn sets up the system, also installs Homebrew, installs apps from the App store and sets up all my configs. The only thing I need to do after is sign into some accounts.
Speaking from the viewpoint of a whole operating system images, the main challenge is that while Nix allows you to create ephemeral environments, many people (myself included) have various hard-coded paths for mounting hard drives. If you want something to be shareable, you have to create a workflow in which the user environment is activated interactively after a tty session is acquired. Same goes for any system services that need persistence -- these have to be configured to be activated at runtime. It's a lot of work for a party-trick. It's probably possible to configure the system such that the log-in needs a FIDO2 key which is also used for LUKS drives, which would be similar to how macOS handles log-ins. But abstracting this such the login works on every machine possible suddenly requires filesystems to be networked, and so on.
That being said, we used NixOS images to boot several Windows PCs of my friends into RAM to play Halo 3 multiplayer split-screen. Most of my friends were mainly confused why they could play with any gamepad they had in their shelf. They also left the event with no permanent changes to their PCs.
I haven't used Nix yet, but I have done exactly this a number of times with Guix now. I assume Nix has the same capabilities. In Guix you can specify a home configuration that includes packages, configuration files, and running services all in one manifest.scm file. If you want to make sure that it's isolated from whatever else is set up on the system you can launch that manifest in a container with a single extra flag
this is basically the purpose of nix flakes, which take in inputs (usually git repositories) and provide outputs (compiled files, docker images, etc). it's a pain to get going, but not tremendously difficult with some googling. It's a lot easier with AI as well.
Installed nix-darwin on 26.0, on a fresh M4 air. I have updated since updated macos to 26.2 through the normal method, no wacky nix stuff there. no issues. no clue on major version changes, but nix-darwin is essentially the nix config language parsed to then run the necessary set of scripts.
I’m using nix to set up both my dev laptop (macOS) but also my self-hosting/homelab (NixOS). It works really well, and nice to have one way to set stuff up. And they are both synced up with the stuff I want to be mirrored across systems.
I see these kind of posts like using Gentoo, cool that some people are having fun, not for me, using computers the mid-80's I have better yaks to shave.
I am critical as with any technology (see my comment about Nix in another thread here), but Nix is profoundly not using computers the mid-80's way. It is more future technology, where a system is immutable, fully declaratively defined, has atomic updates/rollback, etc. It's the direction things are going, see e.g. macOS sealed system volumes, Fedora Silverblue, etc. It's just that it still has a very large number of sharp edges, not so great documentation, lack of static typing, etc.
Nix may not be the tech that replaces everything, but at the very least it is and has been an important exploration vehicle for declarative configuration, immutable systems, etc.
I tried nix-darwin for half a year, and ran into endless problems: poor docs, huge default disk usage, non-trivial to customize, etc. After 6 months, I eventually went back to a mix of Homebrew and mise, which does most of what I need for <1% of the hassle.
I still believe something like nix is the future of building software, I'm just not sure it'll be nix itself.
Interesting, and good luck to OP. I feel a little clickbaited, cause I was hoping for a port of stateless infra (although it may be of little use on a (dev machine :p) [1]
A useful way to frame this isn’t “is it worth tens of hours to avoid a future reinstall” but “where do I want my entropy to live”. You’re going to invest time somewhere: either in a slowly-accumulating pile of invisible state (brew, manual configs, random installers) or in a config that you can diff, review and roll back. The former feels free until you hit some cursed PATH/SSL/toolchain issue at 11pm and realize you’ve been paying that tax all along, just in tiny, forgotten increments.
I think where Nix shines isn’t “one laptop every 6 years” but when your environment needs to be shared or recreated: multiple machines, a team, or a project with nasty native deps. At that point, nix-darwin + dev shells becomes infrastructure, not a hobby. You don’t have to go all-in on “my whole Mac is Nix” either: keep GUI apps and casual tools imperative, and treat Nix as the source of truth for the stuff that actually blocks you from doing work. That hybrid model matches what the article hints at and tends to give you most of the upside without turning your personal laptop into a second job.
One of the biggest annoyances I have with doing this with Nix vs another tool is that Nix doesn't natively communicate back state changes so that you can make them reproducible.
If I make a git repo, place '~/.config/newsapp.conf' in there and then symlink it back to '~/.config/', if NewsApp introduces a new variable in its settings I am immediately aware because Git will complain about being dirty. However, Nix will happily just nuke that .conf and rebuild whatever is in your configuration, without letting you know about state. Which is ultimately bad for reproducibility. It's a huge blind spot in Nix.
It sounds like you're talking about home-manager, which is a third-party Nix module, not Nix itself. I've been using Nix happily for several years now for work and personal without needing to use home-manager at all (since I don't like it either).
I tried getting Nix working a couple of months ago and ditched it because changing some Tower settings updates the global gitconfig (as it should in this circumstance) and Nix would wipe them out.
All of it seemed way too annoying compared to just having a dotfiles repo, and if it couldn’t handle the Tower/gitconfig issue I know for sure everything else I was going to run into wasn’t worth it.
>I think where Nix shines isn’t “one laptop every 6 years” but when your environment needs to be shared or recreated: multiple machines, a team, or a project with nasty native deps.
I'd like to add the third thing, which is just iteration. It's very tricky to maintain advanced workflows even locally. I'd guess many won't even try to compose things that could work in combination (often self-hosted services), when they know they can't reliably maintain those environments.
I've tried nix-darwin a time or two in the past. Every few years when homebrew makes a "hostile" change and I get upset I consider trying it again (now most recently with changes to gatekeeper). I think I'll get to doing so in the next year or so.
But I think just in fairness, the comparison here for flakes should be to Homebrew bundles. My packages are managed in a bundle: https://github.com/Julian/dotfiles/blob/main/Brewfile and then locked by a lockfile: https://github.com/Julian/dotfiles/blob/main/Brewfile.lock.j... and installing is just `brew bundle install`. All native Homebrew functionality. In practice I have never had an issue with non-reproducible builds across my machines (partly because the tendency on macOS is to run the latest versions of things and stay up to date).
(But again I do find nix-darwin interesting to try for other reasons.)
Very nice, I think I'll be moving my "must-have" homebrew packages to a Brewfile. FYI tho, Homebrew no longer supports Brewfile.lock.json (it was always just a log anyway, not a lockfile). https://github.com/Homebrew/homebrew-bundle/pull/1509
> the comparison here for flakes should be to Homebrew bundles.
The bundler integration for nix-darwin actually just bakes tightly-controlled Brewfiles. It’s still worthwhile though, since part of the “tightly-controlled” means better cleanup when you remove things.
I see this:
> The consequence is me, spending a few hours debugging my environment instead of writing code.
But then I also see this:
> I’ve spent a lot of time recently moving my entire workflow into a declarative system using nix.
I can see how this can be beneficial for someone who switches systems very often, reinstalls their OS from scratch very often, or just derives a lot of pleasure/peace of mind knowing that their dev env is immutable.
I change computers once every 6 years or so, maybe more. To me this looks like exchanging a couple (hypothetical) hours of debugging 6 years in the future by tens of (guaranteed) hours trying to climb up the nix learning cliff.
I am happy that it works for the author though, and knowing that it's possible is good in case my particular development circumstances change.
Nix is not worth it if all you want is configuring your home computer. The learning curve is steep and has a tall onboarding cliff.
The only way you get positive ROI from Nix is either you enjoy the journey, or you use it to do more than just managing a single computer: you manage a fleet, you build thin application container images, you bundle all your software, you have devshells, repeatable tests and deploys, etc. It's the same tool for all of them.
It's worth it for me. I learnt and set up my home manager config once in 2024 and I now only occasionally make light tweaks. But I deploy it on every personal computer or remote development servers at my jobs (previous and current). Granted, I'm probably one of the handful of people in my community who uses nix, because it's too arcane - but I've already paid the cost.
Right.
Nix is a wonderful technology. But I would not argue it is practical if you can afford "just fix it when it breaks". A nix setup more/less requires you to pay all the cost up front.
I appreciate putting in the effort now so that I don't have to later for stuff like declarative dev environments. It's really nice to not have to copy-and-paste installation instructions from a README. -- I did like the point: until you've felt what a comfortable design is, you cannot imagine it.
Yeah, I'm a Nix user and fan but I'm still going to be stuck fixing it when it breaks. The biggest thing that I have to remind myself is that reproducible != "just works." Once you get to what you want, you're set but until then and every change after, there's a chance you'll be in the weeds.
At least from my usage, the fact that your configuration is all tied to whether the entirety of nixpkgs-unstable is working can be a real headache. Like recently when CMake was upgraded to 4.0, it took down a healthy handful of packages, which meant you couldn't update anything until things were resolved or you were really fluent in Nix hackery. (I was the former)
Long-time (by now) Nix user here. It's even worse than a few hours of debugging. In my experience it requires continuous maintenance, options tend to be deprecated and moved around on a regular basis (at least in NixOS and home-manager), things like NeoVim break regularly as a result of friction between the immutable world and a lot of plugins expecting a mutable world, etc. I do install my Mac from scratch every 6 months or so (I like a clean system), but with a dotfiles repo and Homebrew, it takes me about an hour to set everything up again, which is far less time than maintaining a Nix-managed system.
For me Nix/NixOS is by far the most effective for deploying servers, development VMs, etc. I do this regularly, so it pays off - I have a familiar environment completely set up in minutes. Another place it pays off, even on macOS, is for development environments, especially if you have to pull in a lot of native dependencies (which can happen in mixed Python + C++ projects, mixed Rust + C/C++ projects, etc.).
Nix shines in all difficult cases, like setting up complex cross-compilation environments, building for old glibc versions, etc. You set it up once and then it's really easy for yourself and other project contributors to get the same build environment.
Like any tool, a good engineer knows when to apply it.
In the six years you are using your computer, do you ever expect to run into versioning issues and conflicts? Homebrew packages conflicting with local packages, something you compile give needs a different python/ruby/node/rust/whatever version that you have locally installed, you want to quickly try out a new package or upgrade without changing your system but have the option of rolling back safely, need to quickly install a database, want to try out a new shell and shell config but don't brick your system and have the option to roll back, etc. Nix gives you all of that and more for a one-time setup cost. Your argument is correct only if you expect to never change anything on your computer for the 6 years. But if I think about how often I have fought with homebrew or some kind of versioning/path/binary conflicts in the past then the investment in nix has paid off exponentially.
It's also about peace of mind like you said. Before nix I sometimes felt anxiety installing or upgrading certain things on my computer. "Will this upgrade break stuff?" - and often it did and I'd have to spend the next few hours debugging. With nix I don't worry about any of that anymore.
That's mostly solved with env managers for python/ruby/node/..., takes at most a few minutes to fully set up and learn, and doesn't get constantly broken by macOS updates.
Even for things like trying out a new shell you can temporarily move the dotfiles somewhere and restore them back and it still takes less time than converting everything to Nix.
But now you’re stuck with Python. Nix enables trivially simple dev environments that are completely heterogenous. This gives you a powerful form of freedom because it literally opens up the entire software universe to your dev environment in a confidence inspiring way. Not to mention things like parameterising anything you use reliably and setting up environment variables, shell scripts, database service whatever you want. Also integrates with tools such as UV really well. Yes, the language is terse and difficult but once you know it, it’s liberating, and makes you a better software developer in my opinion because you now have a high-end full workshop rather than a small toolbox.
Fully spot on, I don't get what is that hard to set a couple of environment variables, and mayby symbolic links, depending on the OS and language being used.
A simple UNIX script or PowerShell utility takes care of it.
None of the ones I have used during the last decades has ever grown to more than like 20 lines of code, minus comments.
This is my feeling too. Nix is a relatively high time investment for a tool that tries to do everything, when you might not need or want everything and using the specific language’s tooling is more than sufficient and quicker. It takes a few minutes to install and do `uv sync`, or `nvm install`, or whatever, on a repository on a new computer, and it just works. Until Nix gets there, and I’m skeptical it will because of the “purist” mindset a lot of people in the community have, it’s hard to justify it.
I think the comparison is "X-as-code", like with Terraform and other tools.
If you just want a throwaway VM, it's straightforward to create one through the UI cloud console. Whereas, terraform is nevertheless still a useful tool to use to manage VMs.
For stuff like installing development dependencies.. it's maybe not difficult to copy-and-paste instructions from a readme, but solutions like devcontainers or Nix's development shells can be useful even if costing more overhead.
> Even for things like trying out a new shell you can temporarily move the dotfiles somewhere and restore them back...
I think the closest mainstream UX for "you can try out this program without having to install it" is running a Docker image.
:) I'd say Nix is second best at everything related to packages.
until you need to start combining things. Docker is conceptually a VM the encapsulates everything nicely, but it ironically doesn't "compose" nearly as well as nix flakes or shells. With Nix you start out with a base env and can trivially extend it hierarchically and jump up and down the tree super easily, and without having to roll your own microservice architecture each time just to get stuff to work together.
Docker OTOH composes whole services nicely: if my project needs a redis cache and postgres instance, I don't have to faff about with local ports, traefik can pick up my web server, and so on. I use a flake to create and lock a local development toolchain, but it's no help in managing the services.
One thing I haven't tried yet is building a container from a flake, which would have obvious benefits for reproducibility. Still don't think it would help with service orchestration though.
> I change computers once every 6 years or so, maybe more. To me this looks like exchanging a couple (hypothetical) hours of debugging 6 years in the future by tens of (guaranteed) hours trying to climb up the nix learning cliff.
yes, it sounds like it's not worth it for you -- you will have to spend a significant amount of time converting your system to do things "the nix way". you can try to do this incrementally, but it's a time sink, and really easy to get stuck bikeshedding instead of doing work.
for me, it feels like a near equal trade-off between debugging nix, or debugging some random env issues that pop up. i know nix, claude code "knows" nix, a lot of other people online know nix. random env issues are random, and yield worse results on google, and frankly are much more frustrating to the point i would rather spend more time with nix than deal with them. maybe a very weird view.
I have experienced a positive return on investment from using Nix Darwin and devenv.sh since getting a new Mac two years ago. Did not spend too much time learning neither.
Also with migration assistant on Mac you basically do not need this in anyway shape or form.
I do this too. It's not for everyone. At this point it's easily been positive ROI for me, but that's after about two years now of maintaining my configs through multiple machines and MacOS upgrades.
I would recommend it only if this type of thing naturally interests you. I can't imagine powering through the initial learning curve if it felt like a frustrating chore.
That said, if having (most of) your machine defined declaratively in a git repository sounds exciting/useful/comfy, then I would encourage you to give it a try. You can start small by just configuring a few programs or options and see how you like it.
I wrote more about my experience here where I also link to my configs: https://bryce.is/writing/code/fully-nix-pilled
What we should be talking about is how to get rid Apple ecosystem all together.
only thing i'd add is mas for mac app store apps you want to ensure are installed but otherwise i run pretty much the same setup.
When i install a fresh macos i have two commands to run - install nix using the determinate systems installer, then apply my nix config.
It's not quite as streamlined as nixos but good enough.
My biggest remaining pain point is dev envs - i've been leaning into adding a flake in each project, so for example i have a single project that's written in scala 2.13, when i cd into that project dir, the correct jvm version, sbt, intellij etc are installed, some useful env vars and shell aliases etc. - that's all great (i haven't felt the need to adopt denenv.sh or flox yet) but i do find myself wanting a devcontainer sandbox workflow more often these days (blame cli coding "agents"), i lean on vscode for that rather than nix so far. In python (where i spend a lot more time) uv loses a lot of value in nix and i don't like that.
The real game changer after I went full nix OS on all my machines and using it on a a few Mac’s is indeed devshells which seemed so annoying to learn but really really are a game changer
I did this for a while but MacOS updates broke Nix often enough that I usually would spent some time every week reinstalling it. I still use Nix for dev environments because it is great but Nix still breaks sometimes.
I also really wanted to like the declarative homebrew configuration but it also often didn’t work as expected for some configurations and had a lot of leaky abstractions that straight up just broke sometimes.
If I ever go back to managing my Mac with nix I would probably just do a home-manager setup and just install most of the applications imperatively.
Given this was using an intel based machine around the time when the switch to arm came so a lot of breakage also stemmed from that.
I still use nix to handle my homelab.
My setup up on my Mac is as follows:
- Orbstack
- NixOS machine run in orbstack
- My whole dev environment is run from this container and is very transportable
- GUI apps are installed on my Mac using the App Store or homebrew etc. but I try to reduce the amount of installed applications
- if I have to install something that I don’t want to install but have to, I try to do it in a UTM machine.
Try dra[1] and a simple shell script in your dotfiles that installs / updates everything in
along with extending the $PATH. Works great for most of my tools (exa, zoxide, bat, jq, etc).1: https://github.com/devmatteini/dra
I’ll check it out thanks :)
I found the Brewfile and a little `setup.sh` to be more than sufficient for getting a new Mac setup.
Ive been using nix-darwin for over a year now after using nixos with flakes for a bit. I now have a singular repo with multiple machine configurations. Nixos for my home server, nix darwin on a macbook air and a nix darwin with a work config. This allows me to have common programs on all machines but also overlay some specialised packages and programs in certain environments. After climbing the initial mountain, its been very satisfying and things just work. My work laptop died recently and I was able to be fully up to speed in a fraction of the time it would have taken me otherwise.
It would be nice to set up immutable config down to the granularity of a git commit, sharing it as a single file, and be able to reproduce that setup on anyone's machine. Instant onboarding and mirrored production (barring secrets).
With the caveat of it not being a portable single file, I find custom image-based Linuxes fill this gap perfectly.
There's of course Fedora Silverblue / Fedora Bootc with https://universal-blue.org/ and https://blue-build.org/ being good examples.
Recent developments have seen the creation of bootc images for non-Fedora distros too, and at this point I've seen quite a few cool arch-bootc custom images, completely customized to the author's desires. See: https://github.com/bootcrew/, https://github.com/tartaria-dev/tartaria
For me, this is the holy grail. Every time I switch laptops, I lift all my config files and such over, but there's always so many system level configs and other things that you have to go and manually fiddle. On top of that, some apps don't really behave well when you just move config files to e.g. under a different username etc. Would be nice if there was a comprehensive solution to this problem, need to try nix-darwin out.
nix-darwin is essentially this. I have a small bootstrap script to install Xcode CLI and Nix, git clone my dotfiles and activate the config. That in turn sets up the system, also installs Homebrew, installs apps from the App store and sets up all my configs. The only thing I need to do after is sign into some accounts.
Speaking from the viewpoint of a whole operating system images, the main challenge is that while Nix allows you to create ephemeral environments, many people (myself included) have various hard-coded paths for mounting hard drives. If you want something to be shareable, you have to create a workflow in which the user environment is activated interactively after a tty session is acquired. Same goes for any system services that need persistence -- these have to be configured to be activated at runtime. It's a lot of work for a party-trick. It's probably possible to configure the system such that the log-in needs a FIDO2 key which is also used for LUKS drives, which would be similar to how macOS handles log-ins. But abstracting this such the login works on every machine possible suddenly requires filesystems to be networked, and so on.
That being said, we used NixOS images to boot several Windows PCs of my friends into RAM to play Halo 3 multiplayer split-screen. Most of my friends were mainly confused why they could play with any gamepad they had in their shelf. They also left the event with no permanent changes to their PCs.
I haven't used Nix yet, but I have done exactly this a number of times with Guix now. I assume Nix has the same capabilities. In Guix you can specify a home configuration that includes packages, configuration files, and running services all in one manifest.scm file. If you want to make sure that it's isolated from whatever else is set up on the system you can launch that manifest in a container with a single extra flag
this is basically the purpose of nix flakes, which take in inputs (usually git repositories) and provide outputs (compiled files, docker images, etc). it's a pain to get going, but not tremendously difficult with some googling. It's a lot easier with AI as well.
How well does nix-darwin survive macOS updates?
Installed nix-darwin on 26.0, on a fresh M4 air. I have updated since updated macos to 26.2 through the normal method, no wacky nix stuff there. no issues. no clue on major version changes, but nix-darwin is essentially the nix config language parsed to then run the necessary set of scripts.
I’m using nix to set up both my dev laptop (macOS) but also my self-hosting/homelab (NixOS). It works really well, and nice to have one way to set stuff up. And they are both synced up with the stuff I want to be mirrored across systems.
I see these kind of posts like using Gentoo, cool that some people are having fun, not for me, using computers the mid-80's I have better yaks to shave.
I am critical as with any technology (see my comment about Nix in another thread here), but Nix is profoundly not using computers the mid-80's way. It is more future technology, where a system is immutable, fully declaratively defined, has atomic updates/rollback, etc. It's the direction things are going, see e.g. macOS sealed system volumes, Fedora Silverblue, etc. It's just that it still has a very large number of sharp edges, not so great documentation, lack of static typing, etc.
Nix may not be the tech that replaces everything, but at the very least it is and has been an important exploration vehicle for declarative configuration, immutable systems, etc.
Immutable systems have zero to do with Nix, a common source of inspiration maybe, that is all.
So I will be doing configuraions instead of work. Thanks but no thanks.
This is just procrastination.
I tried nix-darwin for half a year, and ran into endless problems: poor docs, huge default disk usage, non-trivial to customize, etc. After 6 months, I eventually went back to a mix of Homebrew and mise, which does most of what I need for <1% of the hassle.
I still believe something like nix is the future of building software, I'm just not sure it'll be nix itself.
Interesting, and good luck to OP. I feel a little clickbaited, cause I was hoping for a port of stateless infra (although it may be of little use on a (dev machine :p) [1]
1: https://grahamc.com/blog/erase-your-darlings/