my hot take is that a distro at the end of the day is just a package manager and an install script.

the question i would ask is: what do you want out of your package manager?

this would be an odd question for your given non-Linux user because their package manager before was called “fuck it YOLO”.

i’ll go through some that i use or have used.

apt/rpm/etc: stable. designed for the Linux of the 90s. uptime is king. i end up using these usually out of network effects. some popular package uses it in their Docker file or my VPS only supports these. they’re boring, but that’s on purpose. updates are purposeful and vetted (relatively). these are your Debian, Fedora, Ubuntu, etc flavors.

pacman: seriously just a great design philosophy imo next to the others. rolling release. no versions. fix forward. also with the AUR, you get your “fuck it YOLO” packages back in a way that they can actually be maintained, updated, or nuked as appropriate. if you ever caught yourself waiting for a CUDA release on apt, this is for you. Arch, Manjaro,… i think there’s a ton of vibe coded configs that all pretty much amount to Arch.

nix(/maybe guix?): a software engineer’s package manager. everything is declarative, reproducible, and version controllable. what was that thing i installed last weekend? it’s in the commit log. need 3 different C compilers/graphic drivers/toolchain dependencies installed? handled by the OS, and cleaned up when those versions change. this is my current rig cuz it helps me keep a bunch of machines in sync for my home projects.

Gentoo and LFS and anything else would be a bit much.

i mean, people are making money. sure there are debts owed, but Mark isn’t going home to a 1br rental even after being so “humiliated”. the bag holders are the investors.

i kinda feel sorry for Yann LeCun and his team at FAI. he’s been saying from the beginning that LLMs are not the way forward for AI research as a field, and that has kept Meta from competing in this gold rush effectively. world models, multi-modality, etc are fine and good research but aren’t here to compete with Claude Code or even as effective at being good enough for stuff like search.

not to be too reductionist, but it’s a tough spot when your head of AI is insisting that this tech has no future when there is a literal $T market for it.

the irony is that these “frontier labs” like OpenAI and Anthropic have basically dropped every greenfield research project that doesn’t directly support their “agentic AI” infrastructure (ie endless Markdown prompts + a while loop) while the labs working on the true frontier are seen as failures because they can’t make a bazillion bucks burning megawatts of energy to replace their front end devs.

people have tried.

people predicted the enshittification of GitHub as soon as the acquisition was announced, as you can imagine. now, picture yourself as a dev in that month where a small vocal userbase is reading tea leaves based on Microsoft’s past behavior telling you to move your project, where the best outcome is nothing changes, to a new platform. you have a hundred issues and a dozen PRs in review, and those won’t stop coming in while you are migrating. now you need to mirror your project on GitHub, unless you want to immediately fade into obscurity, because while you’re spending your valuable time making sure everything is setup as it was but now on GitLab (the only realistic alt at the time), issues and PRs are still coming in, and you have to keep your releases updated in GitHub for a while during the migration. you also need to figure out CI/CD on your new platform.

so the ideal—that you can migrate and nothing changes—is a pipe dream. your packaging is now likely totally different; you’re now that snowflake project in the config where i had to figure out how to point to something other than GitHub and waste 30min questioning whether i need your tool at all. you still continue to get PRs and issues through GitHub because of course they didn’t read the README. and there’s tiny friction everywhere. the UI is different, how OAuth is handled is different, the plug and play you got from GitHub Actions is gone, etc etc.

meanwhile for 6 years things are chugging along fine at GitHub: Actions is getting better, Treesitter support, better UI for PRs.

it’s the AI stuff that’s ruining GitHub no doubt. not the AI itself but the culture around it with the “what is our team doing with AI?” nonsense corporate policy. it’s all happened really quickly, and isn’t the “boiled frog” scenario at all really.

Linux was around before GitHub, and wherever we end up as long as we still have our Unix tools like git it’ll be fine.

ideals are great. the perfect is the enemy of the good

deleted by creator

philosophers are in shambles over this comment.

for real tho, people have been trying to define consciousness forever. the problem isn’t that we haven’t tried; it’s that—as demonstrated by your comment—we’ve mostly failed.

for me the only theory that doesn’t depend wholly on magical thinking is panpsychism: everything is conscious; it’s just a matter of degree.

as much as i hate this garbage site, it was referenced in the article: https://x.com/kdaigle/status/2047803291988590609

come back to

is the real joke here. why would anyone come back? the reason this is such a joke is that GitHub has started to fail not just in Actions or Copilot but literally losing commits, ie the core git technology that has been rock solid since before there even was a GitHub. after migrating away for stuff like this they’d literally have to pay me.

it’s Ubuntu dawg. you get what you pay for.

i guess it would be nice, but packages being a few months out of date is pretty normal for Ubuntu, in my experience. i’m not sure what their testing process is like, but part of using something like Ubuntu is stability guarantees. if they felt like the couldn’t do that for newer versions for whatever reason (resource constraints, lack of downstream interest from stakeholders, etc) they’re not necessarily obligated to.

there’s a world of options. this is an LTS distro. use Arch or Nix or whatever if you want the latest packages. i actually switched to NixOS because the CUDA drivers were too new on Arch, and i wanted a better way to pin versions.

or i dunno keep publicly complaining about it until someone does the work for you

guaranteed any company worth more than a handful of salt does not want this. my company would throw a library of books at them for using data in any way that isn’t 100% explicit. for the longest time they blocked me from running Ollama on my laptop cuz the lawyers didn’t understand how neural networks work and thought i was exfiltrating data.

this is only going to hurt companies that probably shouldn’t be using Atlassian products anyway (ie any company with more agility than a boomer era corporate dinosaur)

it’s one of those cases where if you have to ask, you should probably just use systemd. anything else is outdated or a passion project based on some idealism, which i’m all for, but if you’re worried about gaming performance as a primary concern i’d put it out of your mind. for example, i’m an obsessive tinkerer that uses NixOS and Arch before that and i use nushell and Neovim and Hyprland, but i use systemd cuz i don’t see a reason not to. it’s well supported and stable.

i’ve been a big fan of Jujutsu (jj) since adopting it a few weeks ago. things i used to avoid with git like proper rebasing and focused commits become so much easier, in addition to the benefits of conflicts being easier to handle. the learning curve i thought was going to be grueling only took a couple days to get used to, and honestly interop with GitHub and my team’s particular workflow were the hard parts. so not only is it useful, powerful, and becoming more important to my workflow all the time, it’s a joy to use compared to git.

i guess honorable mention to zoxide, which has basically replaced cd for me since it does everything cd does but also keeps a small db of your most commonly visited directories so you can just do z Downloads or z my_project or whatever from any directory

“unhackable” is a bit sensationalized here. the Xbox One is actually a security success story not because it is impossible to hack, but because it’s a rare example of a console that wasn’t hacked in its service lifetime. at the risk of giving praise to Microsoft, the architecture is actually really neat and informed the security features of subsequent Windows releases, ie a hypervisor with sandboxed sub containers (this is why they required TPMs).

(also i’m not agreeing with requiring a TPM for general purpose machines; they make sense on a bespoke hardware platform like a game console)

i bet this hack is nuts, but the blue team deserves some level of kudos

https://youtu.be/U7VwtOrwceo

lots of good reasons to spin up containers for development is my point, especially for an embedded system with exotic system dependencies (compared to Node.js, for example).

Home Assistant is a “simple app” that i have to open ports for.

does that mean they have a good reason? maybe not.

really? i mean, people have done it for friggin vanilla ass Node.js servers. Android projects can have weird dependencies that a container might help solve (NDK etc).

i’m not saying that it’s my preferred way to build, but that wasn’t the question.

i’d vibe code something in Python for this tbh, but i have some expertise in this area already. you could even get some classification going with a YOLO model to help you narrow down the search. it won’t have a GUI unless you count Jupyter notebooks.

normally it’s for syncing across machines, but it is convenient for setting up new machines. i use chezmoi and Nix and some other tools to keep things in sync

i host my dotfiles on GitHub, but any cloud provider or self-hosted git instance will do. otherwise, rsync, scp, or a good old fashioned thumb drive