Learn to host your own services now. Because in the future you might not be able to discover how.
Since folks have asked, I'll elaborate a bit.
I believe Google/Microsoft/OpenAI/et al will increasingly work to strip search results which provide alternatives to their own products, and that includes self-hosted media.
I've seen this when I've attempted to cover using RSS feeds instead of YouTube subscriptions- I have no reason to trust Google to surface Google alternatives in good faith.
There will come a point where you ask internet-oracle-of-choice "how do I self-host a Netflix alternative" and they will intentionally give you bad advice in order to discourage you.
That point is coming sooner rather than later, and we need to train *an entire generation* of internet users how to get out of this trap.
That's *our* work to do, RIGHT NOW.
@vkc yeah, there's also a large group of people who just claim that running anything yourself is "a nightmare" and essentially call people who do a kind of digital Luddite.
It's very discouraging, some part of it is undoubtedly that asking a lot of people who DO run this themselves aren't great at explaining things.
I'm glad you're jumping into the fray!
... I should really finish and publish my guide on how to run your own email server. *Shame*
Yeah I hate that. Those people have a serious lack of imagination, creativity, and community spirit. But here's the thing.
They aren't wrong.
IT IS a massive pain in the ass. I know because I do it. I've been doing it for over 10 years, and now support services for 100+ people.
IMO tutorials and walkthroughs are great because they are part of building a new experience where it can be easier and it can be understandable in a shorter period of time.
But I'm not sure it's enough, I think we also need to take a critical look at, for example, the UI/UX of linux servers, and try to do better.
I agree with what the person said about NixOS and having techie folks create recipes that can be instantiated by others without the same amount of time investment. IMO something like that, plus usability testing, could make a huge difference.
Usability testing is basically impossible without $$$$ investment and business involvement, simply because of how labor-intensive, un-fun, it is, etc. But the good news is it only takes one -- it only takes one group to break through that barrier and produce a gem, and it can be copied the world over.
@forestjohnson @vkc I'm not going to sit here and say that running my own has always been problem free, but I've literally never had a freemail account (and I'm 40), I run my own Nextcloud, and Wordpress.
For my company I run chat, mail, storage, git, kanban, and video calling.
You have to keep up-to-date on things a bit, like running email now is a lot different than when I started like... 20 years ago. But then again, so is operating any other piece of software.
@forestjohnson @vkc I'm not disagreeing with you that something needs to change!
But what hasn't worked out super well in the past are things like webmin or YaST, they tended to break systems if you looked at them wrong.
I don't think the solution is to make it unnecessary to understand the systems, that has never worked.
I think it would be way more valuable to make good resources for people to understand the systems so they can make their own choices about their systems.
The GNU suite of userland applications that we rely on for linux server administration, plus Systemd. They're great, buuuutt... They dont have any affordances, so they are a major no-go for the general public.
I think a replacement is in order. -- something that is readily available on every platform (iOS, Android, Windows, Mac), something that can list processes, list systemd service units, list docker containers.. do all the other CRUD operations on those, all the while offering commonly-legible affordances (not a manual that starts with "how to read this manual", but instead an explore-able UI that "shows and tells" instead of demanding that the user already becomes an expert before they use it)
We won't get anywhere until this kind of thing exists.. People aren't going to, en masse, wake up on day, find a $30,000 gold nugget under their couch to support themselves for the next year, and then think, hmm, you know what I should really spend my time on??? Reading through the linux man pages 10 times.
@forestjohnson I think that's fair to a point. But nobody has come close to doing anything like that.
Netware was kinda easy, in that respect, but it also couldn't do a great deal.
Windows is fine until something goes wrong. Then the difficulty curve becomes a cliff because it won't ever tell you anything useful. If you encounter a problem for the first time on Windows you have to go ask someone.
On Linux, at least, once you have a base set of knowledge you can mostly figure things out.
@forestjohnson That's not to say that things are great now. Not at all.
I'm just trying to say that the problem is large to the point where it seems that nobody has ever even TRIED to fix it.
@hp Yeah, that's what I'm saying, nobody has done it yet. But that doesn't mean its impossible. Windows is absolutely not the way, but I do believe that a well-documented HTTP-based UI for linux, systemd, and docker, could potentially be a home run.
It would have to include the linux installer too, including managing the installation from a phone, so you don't have to plug a kbd and mouse into the server.
@forestjohnson I'm mostly wondering how to preserve the "useful errors that will tell you what went wrong."
I've never seen that done in anything GUI-like, other than just stuffing a log-file into a textbox. And at that point it'd be better to be able to run grep on it, or find all logs around the same time on the system...
How do you give a novice user information like "This violated an SELinux policy" and let them fix it, without making it easy to accidentally allow an exploit to run.
> How do you give a novice user information like "This violated an SELinux policy"
1. disable SELinux
2. If you want to enable SELinux, you have to make a GUI for it -- you have to actually go into the SELinux source code and add the parts that will enable actual usability. Not to create a shitty error message like "This violated an SELinux policy", but to create an error message that contains the word "because".
------------------------
.. nobody has ever even TRIED to fix it.
Nobody ever tried to fix climate change either... But if we don't fix it, it's all over real quick.
I believe in an interpretation of what we observe about the universe that says that "what we observe is generally what was most likely to happen". aka "many worlds"
In a thousand years, the only likely outcome that anyone will be around to observe, is the outcome where we got thru it...
I took a heroic dose of psychedelics and saw the Golden Path, so I'm trying to walk it. Succeed or fail, don't care, at least I tried and did my best. Sue me.
@forestjohnson @vkc What does GNU have to do with anything? I'm very confused.
You mean, text based interfaces?