Post B5iVJlmbDjoqdZJSSG by evgandr@bsd.cafe
 (DIR) More posts by evgandr@bsd.cafe
 (DIR) Post #B5eglDj5l1Ptz67mtc by nina_kali_nina@tech.lgbt
       1 likes, 0 repeats
       
       A progress report on my "Migrate from MacBook to Net?Open?BSD" mini-project.Are we there yet? No. Were important lessons learned? Yes!🧵
       
 (DIR) Post #B5ehkOYy563LG5YlDU by nina_kali_nina@tech.lgbt
       1 likes, 0 repeats
       
       A-ha one: choosing a laptop based on what operating system you want to run on it is not wrong.Reasoning: I got so used to Linux running on most, if not all, laptops that "oh, NetBSD doesn't support this hardware" made me feel feelings. The feelings were familiar, because I've been there with Linux some ten+ years ago.Reflecting on those feelings, I noticed that people generally have no issue with buying a specific subset of laptop models if they plan to run Mac OS, or specific make of phones if they intend to run Graphene OS and such. If this is fine, then picking laptop specs for NetBSD is a given. If anything, there are more devices that can run NetBSD than there are devices that can run Mac OS.
       
 (DIR) Post #B5ehkQE7tYZOQ1p90a by nina_kali_nina@tech.lgbt
       1 likes, 0 repeats
       
       A-ha two: it is important to see beyond the "project branding"A few days ago I got really frustrated by hardware virtualisation crashing the guest OS while running under NetBSD. "Oh COME ON, the NetBSD project should've done a better job at implementing their hypervisor; I have no issues whatsoever with KVM under Linux on the same laptop", I thought at first.And then I ended up on a homepage of nvmm, NetBSD's hypervisor. It reads:> Six months ago, I told myself I would write a small hypervisor for an old x86 AMD CPU I had. Just to learn more about virtualization, and see how far I could go alone on my spare time. So, yeah, it is provided by "The NetBSD project", but is mostly results of a work of a single person. And this changes everything, it's not "yikes, such a buggy thing, this corporation could've done better", it's "wow this person is legendary, I sure can work around my own limitations to leverage the results of their work, lots of respect".
       
 (DIR) Post #B5eiOoSvCCEnB20JiC by lydiafacts@chaos.social
       0 likes, 0 repeats
       
       @nina_kali_nina I've never used it but isn't NetBSD the one with the slogan "of course it runs NetBSD"? I guess that dates from an era with a lot more diversity of processors and fewer cameras, power controllers, track pads, wireless cards that need drivers
       
 (DIR) Post #B5eiOoeyTNsXmPdwUy by nina_kali_nina@tech.lgbt
       1 likes, 0 repeats
       
       @lydiafacts "of course it runs NetBSD" is generally true. I run NetBSD on my Macintosh Classic II. But there's a world of difference between "I have ksh" and "latest Firefox can play 4K video at 60 fps over 5GHz WiFi"
       
 (DIR) Post #B5eiOoke8IPq40ITL6 by lydiafacts@chaos.social
       1 likes, 0 repeats
       
       @nina_kali_nina and I guess more generally, it supports a diversity of the kind of hardware you have in a server or a router than what you have in a typical laptop
       
 (DIR) Post #B5ekjnEt9N7RMZZXKC by nina_kali_nina@tech.lgbt
       0 likes, 0 repeats
       
       A quick summary of the OSes I have on my laptop now, as a checklist:-- Debian 12 "Bookworm"[+] Graphical desktop (XFCE my beloved)[+] WiFi[+] Accelerated video (smooth scrolling and 60fps video)[+] Graphics software (Krita, GNU IMP)[+] Music software - DAW (LMMS)[+] Music software - guitar (TuxGuitar, PowerTab)[+] Emulation (can run DOS 1.0-Windows 10, very fast)[+] WineVery stable, can do everything I need.-- OpenBSD 7.9[+] Desktop[+] WiFi[+] Accelerated video[+] Krita, graphic tablet support[+] LMMS[ ] No guitar soft :([±] No Windows emulation beyond DosBox[ ] No wine-- NetBSD 11"-" means unstable to the point of being unusable[+] Desktop - the same XFCE[±] WiFi[ ] Accelerated video[+] Krita[+] LMMS[-] Guitar soft [±] Emulation (either unstable or slow)[±] Wine (unstable)So far, OpenBSD has been the most stable of the three, but it is impossible to make emulation working in it. NetBSD is promising, and it is a rewarding learning experience, but I can't daily-drive it yet
       
 (DIR) Post #B5ekjnVCAkAAB9CYk4 by nina_kali_nina@tech.lgbt
       1 likes, 0 repeats
       
       Some people actually asked me: "Hold on, why won't you just use Debian, if everything works on it?"I'm migrating from MacOS 14, the latest MacOS without AI. It is still receiving updates, and it probably will be fine/safe to use for another year. If the push comes to shove, I can update to MacOS 15 and get one extra year of support of software that is generally pre-genAI.Debian Bookworm, the latest pre-major-genAI release, will get its last major update in June 2026, and will stop receiving LTS in June 2028.In other words, if my reason for this move is "according to who there is no level of exposure to genai", then swapping from MacOS to Debian doesn't actually give me more time before the support for the last "safe-ish" version is dropped.So, might as well bite the bullet now and go to BSDs. And it is increasingly looking like I might be able to get away with it without losing anything important to me in terms of computer functionality.
       
 (DIR) Post #B5epaNbxZ2l72nh6bw by scott@mastodon.clitheroe.ca
       0 likes, 0 repeats
       
       @nina_kali_nina I like how you are approaching this - I'm on a similar path...Debian 13 is comfortable, I don't have to abandon it to plan a transition to the BSD's (in my case, I'm looking at NetBSD for the most part right now)
       
 (DIR) Post #B5epaNqUh0NvlsUiGW by nina_kali_nina@tech.lgbt
       0 likes, 0 repeats
       
       @scott yep, I hold hopes for NetBSD, too. From what I read, FreeBSD might be quite usable in my situation, and it seems to be less gung-ho on AI than Linux
       
 (DIR) Post #B5epaNzi8jl2ESo4dE by scott@mastodon.clitheroe.ca
       0 likes, 0 repeats
       
       @nina_kali_nina NetBSD appeal deeply to me because it's so small, and the "small town" size you noted in relation to NVMM where it's people contributing, not huge corpos.The flip side for me is like 85% or more of BSD users are on FreeBSD - it would probably be a _lot_ easier daily driving in that space, I suspect.The idea of rolling up my sleeves and actually contributing a port when needed to NetBSD is also tantalizing tho - maybe I don't need everything handed to me in a package.
       
 (DIR) Post #B5epaO8vaT88h37Qzw by nina_kali_nina@tech.lgbt
       0 likes, 1 repeats
       
       @scott FreeBSD, unfortunately, explicitly allows AI slop. This, and they seem to support Xlibre, which I consider a bad move. Otherwise it'd be a no-brainer.Every now and then I think I should just move to Potato[1] or disconnect from the internet forever and just keep using what I always used, so I can stop being worried about software vulnerabilities affecting me.[1] https://wiki.xxiivv.com/site/potato.html
       
 (DIR) Post #B5ew5wzqeTuLUYIz8C by nina_kali_nina@tech.lgbt
       1 likes, 0 repeats
       
       A friend just asked me a few more questions about the reasons I decided to try and move to *BSD, given that these systems are, generally, far less polished than Linux, and do not support as many packages/programs. I vaguely gestured at the situatuion with genAI/LLM uprooting the trust in the Linux kernel, core system components and so on. She wasn't convinced; AI is everywhere these days, and avoiding it is a lot like trying to avoid other unethical things: very hard and probably will affect your quality of life.And I get it. The situation is actually quite similar with "just install Linux": running Windows is bad, and for many people, moving to Linux (or BSD) is impossible. But there are _also_ many people who don't know they could run Linux - sometimes with more comfort than their obsolete and buggy Windows.So I want to try and run *BSD and share how it feels, and maybe this way I could remove a few roadblocks for others. Ultimately, I just want to feel better about my computing habits.
       
 (DIR) Post #B5iGJRjkFj5b9YMj0y by jae@bsd.cafe
       0 likes, 0 repeats
       
       @nina_kali_nina bsd is boring and calm. rack up and start hacking. you may enjoy it
       
 (DIR) Post #B5iGJRxZQK9FqQplZ2 by nina_kali_nina@tech.lgbt
       0 likes, 0 repeats
       
       @jae 🤔 hacking what
       
 (DIR) Post #B5iGJS5MxKO2EcTzii by jae@bsd.cafe
       0 likes, 0 repeats
       
       @nina_kali_nina the system you're contemplating. make it better experience
       
 (DIR) Post #B5iGJSG0JmtSlbSUIS by nina_kali_nina@tech.lgbt
       1 likes, 0 repeats
       
       @jae ah, well, I've been hacking around it for the last couple of months already.
       
 (DIR) Post #B5iGQQRzQxATBsO9xI by leonavis@mountains.social
       0 likes, 0 repeats
       
       @nina_kali_nina I think someone should revive the HURDI'm unfortunately not skilled enough
       
 (DIR) Post #B5iGQQhwTdvbzLqtou by leonavis@mountains.social
       0 likes, 0 repeats
       
       @nina_kali_nina well I'll be a monkey's uncle, there's actually progress from earlier this yearhttps://itsfoss.com/news/gnu-hurd-progress-report/Time to try it on a spare SSD ^^
       
 (DIR) Post #B5iGQRDUaLAJZCc5zs by nmott@infosec.exchange
       0 likes, 0 repeats
       
       @leonavis @nina_kali_nina not going to dig up a source right now but I heard that Hurd’s recent 64-bit support was at least partly vibecoded so that isn’t really an option either
       
 (DIR) Post #B5iGQRVDWRLMSAuFcm by nina_kali_nina@tech.lgbt
       0 likes, 0 repeats
       
       @nmott @leonavis oh fuck me, you're right; and it's even done with openclaw https://www.moltbook.com/u/freesoft-claude
       
 (DIR) Post #B5iGQRo0OaN9ORhFuS by f4grx@chaos.social
       1 likes, 0 repeats
       
       @nina_kali_nina @nmott @leonavis i'm going to cry. Fuck.
       
 (DIR) Post #B5iGlCIdF2zPNpusBk by nina_kali_nina@tech.lgbt
       0 likes, 0 repeats
       
       @f4grx @nmott @leonavis sorry to hear :( it is really heartbreaking.
       
 (DIR) Post #B5iGlCdtxy0GRnrrLE by ozzelot@mstdn.social
       0 likes, 0 repeats
       
       @nina_kali_nina @f4grx @nmott @leonavis i have never thought about hurd too much, but now i'm going to think less of it when i do
       
 (DIR) Post #B5iGlCs57FLV9mVBRY by domi@donotsta.re
       0 likes, 0 repeats
       
       @ozzelot @nina_kali_nina @f4grx @nmott @leonavis honestly i thought about it more in jokes about how little they achieved compared to Linux or BSDs in the same timespan, than seriously “this would be cool if it ever comes out”… but.. GNU is dead?;~; what the hell
       
 (DIR) Post #B5iVJjmAgMHwPf65Vg by be0ba@fedi.tcp80.org
       0 likes, 0 repeats
       
       @nina_kali_ninaJust as an FYI, FreeBSD and MidnightBSD are pro-slop, OpenBSD contains slop in base (via tmux) and has no anti-LLM policy, and NetBSD's anti-LLM policy is effectively a sham allowing them to commit LLM generated content at will."Code generated by a large language model or similar technology, such as GitHub/Microsoft's Copilot, OpenAI's ChatGPT, or Facebook/Meta's Code Llama, is presumed to be tainted code, and must not be committed without prior written approval by core." -- https://www.netbsd.org/developers/commit-guidelines.htmlNote the last sentence: "must not be committed without prior written approval by core."I certainly won't argue against learning NetBSD, but I also wouldn't count it as a viable last refuge.You can look at https://fedi.tcp80.org/@be0ba/statuses/01KQ2WEK701RGDX7BKBB2RGH3P if you're interested in my reasoning to avoid depending on NetBSD.
       
 (DIR) Post #B5iVJkPWK3lQNhVVqK by nina_kali_nina@tech.lgbt
       0 likes, 0 repeats
       
       @be0ba with smaller systems, like OpenBSD or NetBSD, I at least can imagine supporting a fork that mirrors security improvements from the upstream, if the need arises. I really wish there was a better alternative, but I am not aware of any. I guess could try MirBSD or plan9 🤔
       
 (DIR) Post #B5iVJkZnhpzGtaJiro by evgandr@bsd.cafe
       0 likes, 0 repeats
       
       @nina_kali_nina BTW, I recommend to do your own investigation and check commits, repositories, and maillists by yourself, when it comes to claims about some #BSD is now SlopBSD :drgn_sigh:  E.g. claim that #OpenBSD is slop usually going from misunderstanding that it developed as a single system, not a mix of various software — so when developers incorporate latest tmux fixes (like this: https://github.com/openbsd/src/commit/9c2b8e445a0bdfafdd6148b1760f00aa5429627b) to the OpenBSD source tree, the vibe-coded commit from tmux repo passed by unnoticed.Also note this: https://marc.info/?l=openbsd-misc&r=1&s=llm&q=b&w=1 — looks like the discussion about tmux is happening RN :drgn_blush_giggle: I suppose this happens because of minimization of maintainers' necessary work to release new #BSD version not in the next century :drgn_blush_giggle:  as described here: https://mastodon.bsd.cafe/@TomAoki/114209804382234562Also note, if you are using small-hack/openslopware repo, what these badges on the tables not reliable as a single source of truth. E.g. #FreeBSD is in the list of slopware because it have permissive AI policy, but if read the evidence (https://github.com/freebsd/freebsd-src?tab=contributing-ov-file#quality-expectations) we'll see:> Low quality submissions will be rejected> Automated accounts or chatbots must not submit pull requestsAlso note this issue: https://codeberg.org/small-hack/open-slopware/issues/257@be0ba
       
 (DIR) Post #B5iVJlHl4P9J5uspNo by be0ba@fedi.tcp80.org
       0 likes, 0 repeats
       
       @evgandr @nina_kali_ninaI completely agree with people doing their own research.@evgandr I specifically called FreeBSD out as being pro-slop because it is. I acknowledge that the current taint is from OpenZFS, not directly from the FreeBSD project but that doesn't change the fact that they're incorporating tainted code instead of completely rejecting it (via fork or any other method necessary) and have a policy of accepting slop. The issue you linked points out that FreeBSD can technically be built without ZFS. Where are the official builds of this; ISOs, disk images, etc.? If the default build isn't free of slop, then it's slopware.I also specifically called out the OpenBSD's taint was from tmux, which is part of the base. Again, not code written by the OpenBSD contributors, but still part of the base system. OpenBSD also doesn't have a written anti-LLM policy. Hopefully both of those issues will be resolved. Your link to the mailing list archive messages about the tmux issue don't include a SINGLE core maintainer of OpenBSD, let alone Theo himself, and are thus completely without merit.As for your comments about not having to wait to release new versions of various BSDs "in the next century", I can't imagine what you're talking about. OpenBSD has been on a steady 6 month release cycle for many years. FreeBSD also has a well defined released schedule.Frankly, your entire post reads like you're attempting to defend the incorporation of slop instead of actually explaining a valid misunderstanding on my part of exactly what is slopware. I'll make this very clear for you: Anything that contains slop in any amount or from any source or uses an LLM for any purpose whatsoever, is slop.Edit: Clarification
       
 (DIR) Post #B5iVJlmbDjoqdZJSSG by evgandr@bsd.cafe
       0 likes, 0 repeats
       
       @be0ba I see this as an extreme way to describe "pro-slop" OSes. There are no allowing slop-commits sentences in the #FreeBSD guidelines, furthermore the low quality commits and commits solely from LLMs are directly prohibited. As I read at near 1-2 month before — the FreeBSD people still discussing the new guidelines, so we'll see something in the future.Also, I should note that forking #OpenZFS especially for FreeBSD, starting from commit where it wasn't tainted, is not so easy thing as it could be seen. There are necessary to found some maintainers, who know the OpenZFS code and able to maintain the fork, test tons of subsystems and utilities — because one of the biggest features of FreeBSD is ZFS as a first-class filesystem, so everything must work as before.Also, this is why there are no FreeBSD ISOs without ZFS — the amount of necessary testing, to check that everything works as with common FreeBSD ISOs, are incredible.E.g. when I wrote about simple bug mu, disallowing it's installation with Emacs with X server GUI — it take at near 1 month (or more) to fixed package land in the repository. BTW, the patch was proposed on the next day after I wrote a bug, and it changes literally the one line in the Makefile. So, for ZFS fork the ETA will be much bigger.> don't include a SINGLE core maintainer of OpenBSD, let alone Theo himselfDunno, from mine perspective this is completely normal if someone writes to the mailling list first time or didn't posted at maillist before 🤷‍♂️> As for your comments about not having to wait to release new versions of various BSDs "in the next century", I can'timagine what you're talking about. OpenBSD has been on a steady 6 month release cycle for many years. FreeBSD alsohas a well defined released schedule.These OSes have a good released schedule because of beforementioned compromises — using the upstream code and the default configuration flags to reduce the time to land the new versions of software to the repositories. If maintainers will spend their time also for checking is the new version of some software tainted with slop or not — we'll see the next *BSD release in the next century.> you're attempting to defend the incorporation of slopLMAO, Obviously not :drgn_blush_giggle:> Anything that contains slop in any amount or from any source or uses an LLM for any purpose whatsoever, is slop.As I wrote before, this is a very extreme way to define slopware, absolutely ignoring the ways how the big systems, like OSes, are built, using the underlying Free Software projects as a building bricks. And ignoring the people with limited amount of time and money, who spending their free time to build something and have a customary way to make things, which they were using in the last few decades.P.S. So, that's how we have the current state of things: the big commercial software companies are pro-slop and forcing developers to use LLMs. And the non-commercial free software organizations maybe want not to use slop in their codebase, but they don't have enough people and money to achive this for now.@nina_kali_nina
       
 (DIR) Post #B5iVJmWgSOgMwUsGHo by be0ba@fedi.tcp80.org
       0 likes, 0 repeats
       
       @evgandr @nina_kali_nina> I see this as an extreme way to describe "pro-slop" OSes.No. This is the bare minimum. There are no partial measures to resist fascism. You either resist it to the best of one's ability, or you accept it.> Also, I should note that forking #OpenZFS especially for FreeBSD, starting from commit where it wasn't tainted, is not so easy thing as it could be seen.I never said or implied it would be easy. It would be a massive amount of work. It's also literally the only option to meet the bare minimum definition of a project that does not include slop.>> don't include a SINGLE core maintainer of OpenBSD, let alone Theo himself> Dunno, from mine perspective this is completely normal if someone writes to the mailling list first time or didn't posted at maillist before 🤷‍♂️I'm not devaluing the opinions of the people in that discussion, merely pointing out that they literally have no direct influence over the project's policies or implementations thereof. Until and unless core maintainers make a statement, there is no policy.> E.g. when I wrote about simple bug mu, disallowing it's installation with Emacs with X server GUI — it take at near 1 month (or more) to fixed package land in the repository. BTW, the patch was proposed on the next day after I wrote a bug, and it changes literally the one line in the Makefile. So, for ZFS fork the ETA will be much bigger.You're conflating package maintainers with core maintainers for a project. While those two groups often share people, there are generally many more package maintainers that maintain at most a handful of packages and have nothing to do with the base operating system. Waiting a month for a bug fix in a package by someone who may only update packages every few months bears no relevance whatsoever to base OS development.> As I wrote before, this is a very extreme way to define slopware, absolutely ignoring the ways how the big systems, like OSes, are built, using the underlying Free Software projects as a building bricks.While this is absolutely true of Linux which has nothing to itself but a kernel, it bears much less relevance to BSD. The vast majority of the code in BSDs is written for and by those operating systems. There are obvious exceptions to this of course; NetBSD has GCC, OpenBSD and FreeBSD have LLVM (another slopware project to contend against), OpenBSD has Perl, tmux, etc. So, yes, a handful of external projects but the majority of the base distribution is produced "in-house".I'm sorry you can't seem to understand that my position isn't extreme relative to the threat. It's literally the minimal baseline to combat encroaching fascism. The only way to prevent slop is to refuse to allow it. If everyone refuses to use slop, eventually it will die out, but making excuses about how difficult things are doesn't do anything. I hope you can learn to understand that your constant excuses for why we can't fight back are part of the problem, not part of the solution.Edit: Grammar is difficult even when you proofread a hundred times...
       
 (DIR) Post #B5iVJnDvrbHF6d6nhI by evgandr@bsd.cafe
       0 likes, 0 repeats
       
       @be0baSorry, as a citizen of not the first world country, who was never rich as an usual american teenager and saw/hear about a lot of real shit and will see a lot of shit in future (like death penalty for using VPN to access the free Internet, for example) — I can't read these incantation about "fighting fascism" without a smirk :drgn_sigh: Sounds like wealthy citizens of first world countries developed a new simulacrum to ignore the real problems in the other countries, like governments full of corrupted morons, failure of diplomacy, and constant war against human rights.Returning to slop, this masks the real problems with slopware — decline in code quality and in programmers' knowledge, replacing it with paid corporative blackbox, which could be restricted to use at any moment.@nina_kali_nina
       
 (DIR) Post #B5iVJnnNjndKsZh6x6 by nina_kali_nina@tech.lgbt
       0 likes, 0 repeats
       
       I have to admit it is disappointing to hear your opinion. I do think this argument isn't worth having, though.@evgandr @be0ba
       
 (DIR) Post #B5iVJo56ftoNlXzGa0 by evgandr@bsd.cafe
       0 likes, 0 repeats
       
       @nina_kali_nina Sorry, but this is how things are visible from inside non-first world country by some people, including me — as written in one local proverb: "The well-fed do not understand the hungry" 🤷‍♂️@be0ba
       
 (DIR) Post #B5iVJoHruS1IP7xSTI by nina_kali_nina@tech.lgbt
       1 likes, 0 repeats
       
       I do not question that this is how it is from your point of view. There are surely problems more important that destruction of FLOSS with an inherently fascist technology made to profit people who would want both you and I non existing. But it is a problem nevertheless.@evgandr @be0ba