Obviously we all want to avoid enshittified (aggressively monetized) software or at least get our money’s worth. I’m looking at self-hosting software right now and one I’m looking has a pricing page but only for cloud (no other paywalled features) and is open source. I tried looking up future plans and didn’t find much, so it doesn’t seem like it will enshittify. (not related) I had thought about switching to Omnivore for a long time but then they merged with ElevenLabs and the rest is history.
Your good if:
- It’s self-hosted and open source
- It’s self-hosted and closed source, but doesn’t require online check in ( the DAW Reaper, for example)
It’s a toss up if:
- It’s self-hosted and closed source, and requires online check in.
- It’s self-hosted and closed source, and requires a constant internet connection
- It’s cloud, but open source. (you can generally self-host in this instance)
And you’re screwed if:
- It’s cloud and closed source.
Self hosted in this instance means anything that installs and runs completely locally or on your own personal servers. If someone else’s servers are required I would count it as cloud. Phone apps generally count as “cloud” in this instance.
I’d agree with the exception of Rocket.Chat. I hosted an instance for years. Over that time, basic features kept slipping behind a paywall despite being “open source,” and fully self hosted. It’s fully nagware at this point.
I had my whole group of friends and family jump to Matrix a year ago. Instead of basic features becoming paid features… there are no features!
Edit: I re-read your comment. Yes, RC “requires check-in.” My bad.
Revert to an older version?
That might work for a while, but running out of date software seems like a bad idea.
Generally speaking, you’re not wrong. But I’d say it also depends on the use case and your threat environment. If it’s not publicly exposed, I’m generally OK with running an older version of software. In the case of chat software, for instance, it would have to be available only on local network or a private VPN like Tailscale. That can limit its usefulness though.
Keep in mind that none of this applies to stuff for my job. If I’m working on a problem for my company the risk/reward profile is completely different.
How I decide:
- is it critical software?
- can I switch off of it easily?
- are there other options?
- how good/bad are the other options?
if it’s critical software I try to absolutely avoid anything that can be manipulated out of my control. That’s not always possible, or if it is, you have to make major concessions. An example is home automation software. Before I had learned these lessons I used samsung’s smartthings. ST enshittified almost immediately, because it was 1. built by samsung, 2. dependent on cloud services, 3. they removed capabilities from their API making the product worse than it was before. Home Assistant was coming on the scene right about then and I immediately switched. Home automation software is critical infrastructure, it should never be a cloud service (the point is literally local control, any company selling a cloud-only product for a local-only problem is a grift and will enshittify eventually). Switching off of Home Assistant is dead simple. None of my products in my house are HA products, they’re all their own brands, etc. The actions are all controlled by me, as long as a different system provides similar capability then the cost to switch has nothing to do with HA and only to do with the target system.
Compare that with something like Github. I paid for Github before it was free, because the features were worth it to me. I lose nothing by using a cloud service for Git, because it’s a DVCS. All my data is local to begin with, and Github provides me with things like 1. a community, 2. an account to contribute to others, 3. a hosting location for a portfolio, 4. free static website hosting. The cost to migrate off of Github has nothing to do with Github’s product and only to do with the community that uses Github.
You have to make this cost benefit analysis for everything in your life. Essentially, mine is completely based off of how much work it would be to move off of the product and if that work is worth it. Some things are worth it to self host, for example my jellyfin server is an absolute metric fuckton of work and costs a significant amount. But the alternative to that is paying for 15 different streaming services and still not having access to everything, and being screwed when those companies decide they’re not going to have a specific item anymore, or they’re going to charge for it, or they’re going to increase prices without improving the service. Some things are not worth it to self host, for example I pay for Obsidian because I wanted to use it and iOS sandbox doesn’t allow Obsidian to save to a dropbox folder. There are plenty of free alternatives, but I like Obsidian’s feature set and switching off it at a later point is easy, it’s all markdown files.
Interesting, what is it that makes jellyfin a ton of work? I upgrade sometimes when i remember to and beyond that it just runs without touching it unless i want to mess around with trying new beta features.
Sorry, jellyfin isn’t that much work. It’s maintaining the whole thing, radarr, sonarr, etc. all running on an unraid server that needs maintenance and upkeep. Just work.
I could have said unraid server I just thought more people would understand if I said jellyfin.
Foss or it is a toss
