Why don't more people self host / host everything themselves these days in the age of privacy?
How can we make more people self host their data rather than giving it to corporations?
A start might be to tell people to use extensions that are adblockers and to disable javascript on websites and even use and setup pi-holes to take back their data and privacy.
There must be more that can be done here but it is a start!
Because it's a pain in the ass and people don't care enough to. People don't want the fediverse, they want an app. They don't want to know about cloud and infra primitives, they want their photos in their phone and for them to be safe and shareable. They want their iMessages and WhatsApp to JustWork(TM).
> How can we make more people self host their data rather than giving it to corporations?
You cannot. Make non profits to manage this infra. Wikipedia, Lets Encrypt, etc. Hacker News is a benevolent dictatorship funded by a VC fund and run by someone passionate about the work. Incorporate a 501c3, spin up a deposit account, and start sourcing donations and funding to pay for infra and people.
This is not a doomer comment. Enable the power admins and passionate folks to deliver a great experience to normies with as little effort as possible. Literally how Wikipedia functions: datacenter, paid admins, passionate volunteers curating bytes.
(Would be cool if Cloudflare supported one click deploy of Fediverse tools into their edge cloud; scale when you need, get your data out of R2 and databases on demand so you're not locked in; having to pay for infra is inevitable, that is to be solved for)
- power's out at your house. You should have bought a UPS.
- internet's out at your house. You should have a router with a backup connection.
- security exploit in the wild. You should have configured automatic updates.
- update repo changed URLs. You need to change the config.
- you enabled open signups, and now strangers are distributing MP3s from your images directory. You need to kill those accounts and change the config.
- you let your cousin have an account, and now he's in a flamewar with half of Australia (which he thinks is Austria, and also he believes is ruled by the Illuminati). You need to shut that down and talk to other admins about getting off of their blocklists...
- Your ISP is [ISPs are, to your 2nd point] actively hostile to running "servers" from your connection, so you must either pay a ridiculous premium for that privilege, or jump through hoops to evade their intentional breakage.
- Your other cousin does something illegal (sells drugs, posts revenge porn, threatens a public official) using your host and now the police are knocking down your door in the middle of the night and dragging you in for questioning. Even if you avoid charges, your neighbors eye you suspiciously from then on.
Everything is a possible point of failure. Of those there are those you can control and those you can not. Just adding in a DNS resolver ups the number of possible points of failure by 2. Mix in a proxy server with TLS rewriting. Add in a few more. Add in your docker source of containers is gone more failure. That on top of your usual 'computers are broken in weird ways' most of the time.
Outsourcing that to a 3rd party is tempting, very tempting. But you are also sacrificing other things to do so. So you have to balance those two opposing forces. Sometimes picking up the phone and saying to someone on the other end 'fix it' and they fix it is useful. Other times you digging thru hundreds of forums (or chatgpt these days) and figuring out what is wrong is interesting too and has its uses.
That most people just sign up for something and just want it to work. I totally get.
It seems that I enjoy the pain with many other people on r/selfhosted. I will call that pleasure I feel, The Pleasure of Pain.
I think that was what the web was about since the beginning and lasted for a good couple of years before targeted ads and "data became the new oil".
Edit: I replied before the GP edited the comment.
If you're not participating in https://lemmy.world/c/selfhosted from your selfhosted Lemmy/Kbin/Mastodon/etc... instance, I question your commitment.
Lets Encrypt serves nearly 192M websites with 13 full time staff and an annual budget of approximately $3.35M, for example.
I'm not worried about who owns my data, but I do want to have a stable identity. The first Mastodon server I signed up with mysteriously disappeared after a couple of weeks, so I decided I was better off paying a small monthly fee to put it on my own domain.
Here's my pitch...
Sell a productized version of a server that has everything you need to run all of your data-sharing needs already set up with a nice front end that can be operated by a remote control from an HDMI-connected TV. Using that front end, the user connects to the local network, establishes mobile app companions, enters in any global details for accounts they want to maintain, and manages all the configuration options for the server.
The server would host all of the things a household would want to maintain, using open source projects for transparency and maintainability. That would include things like peertube and mastadon for publishing content and media, but it would also include home automation software, as well as personal media software like owncloud as a way to replace google drive content and plex to manage personal media playback.
Basically, a little server that uses open-source software to emulate every modern cloud-based service, on a household scale so that you can run it cheaply enough to be affordable ($60-$70 bucks?), layered in encryption and firewalls for privacy and federated to other home servers (and everywhere else) using the fediverse, while also adding in anything you would want a home-management or home media server to do.
I call it an "accent server". Like an "accent table". I would make it stylish enough to display, but discreet enough to tuck out of sight.
And, personally, I see this kind of thing coming around either way. It's just a matter of whether or not one company puts together all of this software and starts offering it as a walled garden, or if it bumbles itself together out of CLI utility chaining, and enough reddit posts circling around the same setup questions.
That assumption is based on the idea that most people seem to want what this would provide; it's just that not even particularly tech-minded people want to go through the steps of setting each of all of these things up. And it's only when you have 3 or 4 of the services or features working in tandem that they add up enough to make a change in lifestyle (which is what we're attempting) tempting enough.
So if you could put together a "buy it once, plug it in, set it up, forget about it" kind of offering, I think you would get a ton of people that would buy it, and then once it was just a thing in your house that you could start adding custom plugins to (as easy as installing an app), then you would get a ton of adoption. The hard part is marketing; you have to really explain it fully to make anyone understand what it is you're trying to offer.
- The raw idea seems easy.
- The initial implementation seems like it should be of moderate difficulty, but is actually very challenging to get even close to right.
- The long term maintenance is a nightmare, but don't worry, you won't survive long enough to worry about that.
- The infrastructure and policy implications of getting and keeping it connected to everyone else are intractable. (See https://news.ycombinator.com/item?id=38531969 for some tip-of-the-iceberg examples.)
And yeah, none of that even touches on marketing.
I will say that, for the most part, these companies try a kind of "lock you in to our product which uses open source" scheme that could never possibly work. And, further, that no one has ever implemented the kind of system that I have in mind that I've seen. But that's not because it's unique or complex, just that it isn't a good path to a minimum viable product, so it isn't a good way to spin up a company quickly.
But yeah; aside from burning through cash in order to build enough coverage (maybe a year of dev just on this; no product dev yet) for a product that you will never actually profit from, I don't see how anyone could bring something like this to market.
All of that aside, a product that can sustain a company is not the only way to have a product exist. Modular productization and loss leading are a couple of ways to envision this. But I'm betting some kind of fractional componentization starts happening that makes this kind of stuff more maintainable. YMMV, though!
I might also self-host if residential ISPs didn't treat home servers as an aberration.