Fuck the cloud (2009)
ascii.textfiles.com
ascii.textfiles.com
2020s, people are going offline to have fun.
Homelab is becoming a thing even for people who never had experience with computer, people hosting their own documents, movies, music, backups in case things go bad.
Even some companies have realised the price of going cloud, some are moving back to on-prem hardware with full control.
Oh totally. I got my brother, who lives on the other side of the world and who's not a dev/sysadmin, just a poweruser, to install Proxmox and he's now using GPU passthrough to have VMs run different AI models (in either Linux or Windows) for image generation, experimenting, etc. He's also got a NAS with RAID etc.
To me a homelab is the 2020's version of having fun with computing: there's something incredibly refreshing in disconnecting my sub-LAN from the Internet and still have music, movies, private pastebin (yup I use this at times between computers for simple stuff I don't want to both scp'ing), private Git repositories, complete backup system (including offline HDDs/SSDs that I rotate into a safe at the bank), etc.
A movie projector, a dumb one, is another very cool thing: connected to nothing but a HDMI cable (not that HDMI is the best standard ever but it does the job).
And to be sure I can still code and work without having a nanny holding my hand as if I was a toddler, I regularly have coding sessions where I don't use Claude Code (but I also pay for a subscription: these things aren't mutually exclusive).
For anyone who wants to have a fun, a used HP Workstation with ECC memory is basically $200 and makes a perfectly fine server at home. Doesn't need to be up 24/7 either: my online service that is up 24/7 is my unbound DNS resolver and I run that one on a Raspberry Pi (for the low power consumption). The rest of my homelab (two Proxmox servers) is basically something I only need when I'm awake/at my desk. So I turn them off at night.
You never go full cloud.
I guess I'm just old though.
It was goddam glorious.
Took until 1995ish to have a homelab to experiment with FreeBSD and later Linux over a 10-Base-T network with gcc/g++ and dialup access to this thing called "The World Wide Web". The browser had a throbber dinosaur.
It was even more goddam glorious.
Right now I've got three main systems with decent CPUs and 128GB of memory, and several emphemeral satellite systems. With 8GB of NVIDIA VRAM I'm running gemma4:31b just fine on my media system. Which curiously enough has, ah... media on it.
I feel like I have a good idea how EV owners feel right now. (We have a Prius.)
The difference is that you don't own your EV, it is a computer on wheel. Any firmware update like Tesla has done in the past and features are no longer available.
That is totally opposite of homelab, you have full control, you flash firmware that gives you full control over devices.
I am hard core Linux user, my wireless access point runs Linux, my router is a Sophos baremetal running OPNSense/FreeBSD Unix. My 3D printer is DIY running Debian Linux.
That is the best thing about homelab, nobody can take it away from you, you own everything, it is yours and yours only.
Yeah, I think that's right.
I only thought in one dimension: reliance on corporate controlled high density existing infiltration of fossil-fuel delivery infrastructure. Which is worthless if the price is occasionally exorbitantly volatile or might even run into zero supply issues.
Another equally important dimension is: that EV car might just be a puppet, and not you running the puppet.
I'm pretty sure the Prius doesn't phone home (2015), but I admit that I've not gone deep into it.
I can't stand this thing I just did in this comment where I tried not to sound like an AI. I might have to give up short comments entirely because I can't generate enough context for authenticity credibility. <= It's a fact, and that right there sounds like AI to me now.
You could unplug or shield the cellular modem to take the car "offline."
Make it go offline and watch things going south haha
My excuse is that I never had the financial stability that I have now in my middle 30s to get things going, also moving oversea and what not didn't help either.
But I didn't go crazy, I have 3 Proxmox servers running a few services, Pihole + Unbound as recursive DNS to avoid DNS poisoning and personal data tracking.
A DIY TrueNAS as the primary system to have a copy of my data.
I have a 4K bluray with physical media, but I do have Jellyfin also because nothing matches 80s, 90s, early 2000s movies and buying DVD in 2026 is pointless. Also, it is not easy or very, very expensive to find a bluray copy of old movies in 2026. Jellyfin solves that.
All my servers are consuming 110W 200VA tops, connected to a second hand APC UPS 1000VA.
If the whole world goes to shit right now, I can still run all my stuff without dependency to the internet.
My last goal is to have a solar/battery system so if WW3 really happens sending us to the cave age, wherever I am will still be 21st century.
He wouldn't have paid for it either way, or would have either way due to taxes and how a lot of libraries work.
And it wouldn't have taken anything away from the library itself or other library customers.
We're old.
A lot of HNers weren't born yet.
dude this is way more than "power user" you're being unserious.
If you tell a genuine power user, someone comfortable with Windows registry edits, Office macros, maybe some light PowerShell scripting, that they can "totally do what my brother did," and then the actual task list is Proxmox installation, IOMMU group isolation, VFIO stub drivers, GPU passthrough debugging, RAID configuration, and multi-OS VM management, subnetting, raid and HBA configuration, you're setting them up for a brutal wall of frustration.
PSA: the answer to "you're being unserious" starts with "are you fucking kidding me?"
> He's also got a NAS with RAID etc.
Does the term "hosting" come from "web hosting" or some earlier terminology
Does the term "hosting" in the "homelab" context mean storing data locally on own computers, or running locally stored programs
If yes, could the the term "storing" be used instead
If no, then why is "hosting" the term used
This is sort of rhetorical question. I think I know why but I'm looking for clarification
Not entirely unlike how viruses (or memes) execute on a host organism ;-)
This sounds like a filter bubble plus wishful thinking. Most people can barely manage their phone settings, let alone run a homelab.
Whet I hear alot about is regulation, which is a bit new, particularly with social media platforms. I know quite a few folks who think they should have less ability to arbitrarily ban accounts for example.
While that may not be the most common sentiment other I think about that in the greater context of these conversations I’ve had with other groups and the trend seems to be that nobody wants to replace Instagram per se, they want better stewards and better regulation on social media.
The traffic on every social media is increasing.
There is no mass creation of homelabs and no this will also not be “The Year of Linux on the Desktop”
Also “some companies” is amorphous. All of the cloud providers announced record revenue growth gated only by a lack of being able to get chips
There was a comment in another post on the front page about how anyone "remotely technical" can set up a docker container, and I think this is a good example because the mechanics of it are simple (edit a couple text files, run a couple commands), but half the world couldn't tell you what a terminal is and they're focused on other things in life instead of learning how computers work. Cloud succeeded because cloud is easy (at least in the beginning), it's that simple.
If we are to solve this problem, we're going to have to make self-hosting easy enough for the average 7-8 year old to do it without struggling. One promising way forward is with local-first E2EE sync and backup. The only good implementation I know of personally is Obsidian Sync, which has a UX that I adore, and hope to see more of in the future. There's other good options too, but none that I'd feel comfortable trusting a seven-year-old to execute correctly first try.
It used to be easy enough in the 90s, when plenty of folks had their own custom website. You signed up with a hosting company; they provided you with a bunch of different ways to upload files; your website was hosted.
Of late, I've decided that the problem is that HTML development halted at what is still a very beta product. HTMX is a reasonable attempt to continue the spirit of HTML; where I'm going with this is that I think HTML should have continued to add enough reasonable features, without needing Javascript or massive amounts of CSS, so that most websites most people wanted to develop would still be straightforward enough to do. HTML stopped before it even had a usable <table> with sorting and filtering defined, so we've spent decades with inferior tables in every web app that all suck compared to what we got used to with even Windows 3.1 apps. HTML finally grew date and colour pickers but it should have had all kinds of rich UI controls and behaviour that would have made it totally unnecessary for people to build all the Javascript middleware that essentially treats the browser as a display canvas and otherwise totally reimplements the GUI from scratch. And we wonder why the beautiful new Macbook Neo is kneecapped by only having 8GB????
It's time for HTML6. My standard will be: everything a restaurant website needs should be basically batteries included, with the exception of an e-commerce backend. It should all be doable in static HTML files with almost no Javascript and really just enough CSS to set artistic theming elements instead of having to do arcane shit just to position things.
But that is not self-hosting. You're still using a cloud service. The problem is how to run something local, at home, that you have full control over
Thankfully, the converse: the computers these days are focused on nothing else but learning how humans work.
Hell, half the world doesn't even have a computer with a physical keyboard.
A distinction worth making is between "self-hosting" (running docker-compose, Proxmox, etc.) and "local-first software" (applications that store data on your own machine with no cloud infrastructure required). The former is hard, the latter is just how desktop software worked before SaaS took over.
In small business software the shift has been nearly total. Tools aimed at craft makers, small food producers, etc. have almost universally migrated to monthly subscriptions. The practical result: you're paying $tens-$hundreds/month to track whether you have enough beeswax for your next candle batch, the price increases annually, and if the vendor folds you get 90 days to export your data (if you're lucky).
These users won't set up a homelab, but a desktop app that installs normally, stores data locally, works offline, and has a one-time price is achievable - I've been building one [1] and it's a reasonable middle ground between "trust us with your data forever" and "configure your own NAS."
[1] https://kitted.site (inventory and production management for small manufacturers)
Obsidian Sync gets around a major platform problem with Apple iOS devices, which is that they don't allow one app to change the data of another. You can use Syncthing for local E2EE sync, but it won't work on your mobile due to this. It works fantastic machine to machine. I'm paying for Obsidian Sync now just to get around that, but it shows how some of the platforms are made to prevent functionality. Ostensibly its for security, but I'd argue the benefit is mainly financial for app makers (and therefore their app store cut).
> Resource Limit Is Reached
> The website is temporarily unable to service your request as it exceeded resource limit. Please try again later.
Yeah I get the impression that making money from advertising isn't the primary reason that site exists. I know, hard to fathom right?
You would say that if you look into my 12U rack right now, only 6 months ago all I had was 2x Dell SFF second hand computer from eBay that might have costed me AUD300.
Before that, I had one of those miniPC with two network ports that cost me AUD200. I installed Proxmox in it, then OPNSense (router) and pihole as virtual machine, it ran like that for years
Install Proxmox in them and you can run eveything.
This is the major misconception regarding homelab, you absolutely don't need expensive gear. A single miniPC + Proxmox is all you need to start, try to have at least 16GB of memory, 256GB NVMe is more than enough to start.
Don't let those massive homelab setup you see on the internet tell you that is the only way :)
An inspired nerd can do it right now, but grandma will be able to do a curated, accessible set of things by the end of the year, and by the end of 2027, the internet and self hosted things are going to be incredibly different. When people can self host plex and anonymously pirate anything, and their local model can do the ethically gray area stuff like ensure everything is done so they don't get caught - cloud services can't compete with that. Cable and netflix and spotify and the rest are going to have to up their game, and not do the stupid lashing out, price gouging, hunting the pirates type of thing or they're just going to burn down faster.
We're headed for some really cool, interesting times.
Just get an old server for free somewhere and go …
Well that's your problem right there. The home labber setups are for experimentation or "hot rodding" purposes and they typically way overbuild their solutions.
What most people need is an old desktop in a corner somewhere (preferably close to your router so you can get to it with an ethernet cable).
It's won't be Grandma proof, but if you're remotely technical you can write a docker compose file that glues together some useful home server utilities that sound interesting to you.
My setup is roughly speaking: Ubuntu LTS, ZFS (with 4 disks in a RAID10 style config), and a docker compose file that runs plex, transmission, syncthing, vaultwarden behind an nginx-proxy[1] container that even automagically renews my Let's Encrypt certs for me (though it's probably even easier if you use a Cloudflare tunnel).
If you're confident all your apps are available on these platforms, the storage part is easier with something like TruNAS or Unraid. If you don't need storage at all you can slim down your hardware a lot and just use a raspberry pi.
IMO, just find an old beater machine and get hacking :)
The simpler way to go on most fronts is some form of Proxmox with things like the above managed, it takes care of much of the overhead and doubt on it's own or through a reasonably point and click interface, which could be pre-configured.
Homelab = Experimenting with environments you might use at work. Selfhost = Hosting what you need at home.
These are two very different goals with very different reasonable choices. People homelab with Kubernetes clusters, selfhosting with Kubernetes is dumb.
Gambling and endless consolidation feel good for monkey brains. Governments are supposed to step in, but we have a heckin' Cheeto in the White House.
Fuck the Cloud (2009) - https://news.ycombinator.com/item?id=10771539 - Dec 2015 (219 comments)
Fuck the cloud (2009) - https://news.ycombinator.com/item?id=2984083 - Sept 2011 (2 comments)
Fuck the Cloud - https://news.ycombinator.com/item?id=441885 - Jan 2009 (23 comments)
I mean, it was 2009. How much of your personal data from then is still around on non-archival media you still control? Even among the geek set here, the answer is likely to be "almost none of it". At best it's "backed up" on media you haven't validated.
Or more likely, copied somewhere else to keep it secured. Like... Dropbox or Backblaze or S3 one of those, you guessed it, CLOUD services.
Likewise, do you still have your email from 2009 online in a useful form? Gmail users, many of them in this very thread, still do.
I think the drive that held my old home directory might have died, though.
The primary exception would be Google Photos pictures which were auto-uploaded from my phone that I haven't curated and downloaded yet.
I predict I will maintain my custom-domain email address much longer than if I had used Gmail, given the attrition rate of bannings without support.
> on non-archival media you still control [...] Or more likely, copied somewhere else to keep it secured.
Hold up, is this OR or XOR? It sounds like you're trying to add unreasonable (dis-)qualifiers. TFA isn't saying one must boycott "the cloud" and erase all data, it just advocates that you retain an independent copy.
> Dropbox or Backblaze or S3 one of those, you guessed it, CLOUD services.
I think that's conflating different use-cases.
* Having a regular offsite backup into S3 isn't that different from when the data was rsync'ed to a Linux machine I paid for an account on. Any cloud-ness is a remote implementation detail, not a change in the consumer relationship.
* In contrast, "all my photos are in the cloud and my friends and family can collaborate on shared albums" is different, it permanently moves the locus of control.
No, it doesn't. You're fooling yourself. All the criticism of "cloud" providers is predicated on a presumption of bad faith on the part of the provider. Do the same to Amazon and Dropbox and you get the same risk. More actually, since you're not just storing photos but raw backups that might end up with chat logs or password or authentication tokens or whatever.
All you're saying is that you trust party A but not party B to give you the same service. Which is fine, your trust is yours to give. But it's not an indictment of the technology behind the service!
> Don’t trust the Cloud to safekeep this stuff. Hell yeah, use the Cloud, blow whatever you want into the Cloud. The Internet’s a big copy machine, as they say. Blow copies into the Cloud. But please: (1) Don’t blow anything into the Cloud that you don’t have a personal copy of.
____________
Here's an analogy for how I feel things are going. Keep in mind the differences between: (1) a kind of product-offering, (2) the people offering it, and (3) an underlying set of technologies that could be in multiple products.
* Alt-TFA: "Fuck Asbestos - Everyone's selling asbestos pillows which are dangerous and being pushed by amoral sociopaths. Don't use them without a respirator."
* Alt-ajross: "All your criticism of asbestos is predicated on a presumption of bad faith by the providers. Stop being mean to asbestos. Asbestos can be useful."
* Alt-Terr_: "All asbestos pillows are still terrible no matter who's selling them."
____________
> All you're saying is that you trust party A but not party B to give you the same service.
No, applying logic I choose is a fundamentally different service than accepting data into logic they choose.
No, I was arguing with you, who posited that the difference between Dropbox and iCloud Photos was the "locus of control" and the "change in consumer relationship". That's not an argument about data reliability, it's an argument from trust. And it didn't make sense to me.
Similarly (also from ~2003), the (Australian) ABC's website held a lot of recorded breakfast radio show clips from when Adam & Wil hosted it, getting the awesome comedy band Tripod [0] to write songs in an hour. Many of these were released on their CD's, but nowhere near all of them.
Eventually that ABC server was shutdown due to lack of government funds. There's a very good chance I'm the only one on the planet with these excellent songs & interviews from those shows.
I'm not sure HN is the best place for such... technological anachronistic skepticism? A lot of us ARE going to be storing all that for shits and giggles.
I wish I kept more, honestly. It’s a beautiful record.
I think my most treasured possession is videos of myself and my parents from when I was young. I’m thinking of sitting my sisters kids down in front of a camera for 15 minutes and getting them to talk about their life. It’s beautiful to rewatch this stuff decades later. It’s transporting.
As other commenters have stated, maybe this isn't the best place to ask.
I'm definitely in the "almost all of it" camp. I have Diablo II game saves on my desktop that are carried forward directly from my Windows 98 SE box circa 2002-2003. As well as Linux ISO's I acquired on Kazaa while still on dial-up internet.
Most of my media is backed up on my Unraid server, the most important stuff is backed up on an external drive and I also have some things that exist in the cloud, which I do not trust, which is why it's tiered as least important.
I'm amazed at the number of people jumping out here to insist that people don't use or value cloud storage because of the existance of one or thirty or whatever kludgey manual solutions. I mean, I know you can store stuff manually. I still have all that junk too! It's fun. But I don't recommend it to friends or coworkers or family or anyone else because... well, duh, as it were.
This forum's cherished (and, apparently, deeply insecure) geek cred notwithstanding, THE MARKET walked straight into the arms of the cloud, and has derived immense value from it. Grandmothers have terabyte archives of their progeny's development and will take it to the grave, without needing to puzzle out (sigh) an unraid install.
Cloud services, like everything else in control of rent seeking companies, are getting worse. That was always the obvious, inevitable trap with all of this, with any system where you pay a subscription for remote access to a timeshare computer. Which isn't to say that it isn't useful, I even use it, but I don't rely on it.
You didn't frame your initial post around the market of grandmas, your rhetoric was targeted to those reading your post; "How much of your personal data", "do you still have your email".
Uh... that's wildly and seemingly deliberately mischaracterizing what I wrote. Seriously? The very next sentence falsifies your interpretation, quite explicitly. Why would you cherry pick like that?
Given that, bringing up the needs of grandma and her family photos is a non-sequitur. We clearly aren't talking about that, more considering the wider effects of a tech industry having centralized control of and gating access to customer data and processing and the ways that has caused a lot of exploitation and enshittification.
But clearly you're avoiding talking about any of that, which is why the only thing in my post you engaged with is a bunch of handwringing that I misrepresented something you said (I didn't).
You aren't. I was. Go. Read. What. I. Wrote.
Strawmanning is bad form and bad faith. Please don't do it.
Anyway, I love how well GDPR demonstrated this:
> Insult, berate and make fun of any company that offers you something like a “sharing” site that makes you push stuff in that you can’t make copies out of or which you can’t export stuff out of. They will burble about technology issues. They are fucking lying. They might go off further about business models. They are fucking stupid.
Or if they had a ton of viable competition.
It's an encrustified open source offering where the original vendors aren't compensated. Where there's lock in, proprietary offering creep, and highway robbery billing.
I will not, because that is not what I asked.
>Is not most cloud tech *based* on open-source?
The proprietary pieces are 99% of the offering. It just wouldn't have ever worked without the open source bits.
See e.g. https://aphyr.com/tags/the-future-of-everything-is-lies-i-gu...
(That whole series is also available as PDF; I recommend printing it, spiral-binding the printout kinda like a pirated college textbook, and surreptitiously leaving it in cafes and the like.)