have you considered fucking not?

If you don’t trust the guy who “literally invented rsync in 1996” to do right by his project which he has been doing quite all right with for the past 28 years, then I don’t know how to help you. I’d like to put forward an insane idea that he might know what he’s doing and is not vibecoding with a blindfold on and a beer in his hand.

You literally can keep everything local, encrypted and air-gapped, without it ever seeing outside world.

OK, but I cannot trust the gov to not make it illegal for the average person to use real encryption (even on one’s own hardware)

A law like that would be unenforceable. Nobody knows what’s running in your closet save for police raiding your house (at which time they would also gain access to your physical photos)

Immich on my self-hosted server at home + backups to https://filen.io/

That final boss in Anachronox straight up sucks. I don’t remember if I cheated or gave up and watched the ending on YouTube

same to you!

1. Yes and it all does not matter in the slightest for discussing if a language is a right choice for something like Piefed. Not that there’s much to engage with in the original post - I almost regret leaving a comment on a post like this, but I feel this may be a learning for some.
2. There is zero significance in it being supported in “that” way. There used to be a Go-specific runtime (before there were OS runtimes), but they have dropped it because it made no sense once OS runtimes appeared. If you use TS you likely already have a build step somewhere before deployment unless your function is dead simple (which admittedly it should be, but rarely is). Might as well compile a binary, it’s not that complicated. And yes, using a compiled language like go (or C, yes) may absolutely be the correct choice depending on what you need - if it’s in a hot path which serves a lot of traffic then it may lead to much better warm up times and better throughput, as there isn’t a whole runtime to boot and script to interpret like there would be with Node and/or Python (let’s not talk about Java). I’m simplifying a bit, but hope that helps.

I’m just making a case of how not obsolete python is.

There’s no need to do that, it’s self-evident and the original post does not deserve a second of anyone’s time.

Guess what is and what isn’t used in serverless functions, Lambda doesn’t even support Go ffs. It’s java, ruby, node and python. And from those most use node or python.

I assume that this isn’t bait, so I’ll answer.

1. It doesn’t matter what Lambda supports
2. There isn’t a dedicated runtime for Go because it doesn’t need a runtime. It will work on OS-only runtime like any other language that gets compiled down to a binary: https://docs.aws.amazon.com/lambda/latest/dg/runtimes-provided.html

there may be better alternatives, but migrating the WHOLE WORLD to a new system just because it’s slightly better is not gonna happen. hence - unsurprising.

the fact that imperial unit system still exists IS surprising though

I just realized this almost reads like a Terramaster ad, that’s not my goal; you can search for similar options from other fabricants.

I don’t mind that, I’m actively looking for recommendations. I also got pointed to https://www.minisforum.com/products/n5-air, but that one is 500 with no RAM at all, which is quite hefty.

So far I’m leaning towards trying to assemble something generic from used components, though I’m not quite sure what to pick best, motherboard being the hardest one.

As for TrueNAS - I might check it out if I do get off the shelf NAS, but I don’t imagine I’ll be using the Web UI much, to be honest. And I know about Jellyfin, I’m sticking to Plex just because I already have a lifetime Plex Pass and I like the app on my TV. If they keep shoving “community” features down my throat though - I’ll do the switch.

ECC is not a hard requirement for me, but if I can get it - I’ll try to, as to me it makes sense for something that runs 24/7 and handles my personal data.

I have a very strong aversion to separating storage from my server. I just don’t see why I need to route power and network to 2 small boxes (none of which would do what I need it to do on its own + considering very crappy room layouts in rented apartments) and then fiddle with network access, when 1 slightly bigger box would do what I need it to do. Some 7-8 years ago I’ve bought dirt cheap second-hand Huananzhi x79 with Xeon E5 and DDR3-ECC with some low profile NVIDIA GPU and it all still works now - and something like that would mostly be OK for me even now (except I left it in another country).

That said, it’s possible a reasonably powerful NAS will be enough for me on its own?

storage box and a services box

That’s exactly what I want to avoid though. I see no reason to power and network 2 different small boxes when just one slightly bigger one will do. And as mentioned - 500 is without HDDs, I plan to use the ones I have for now and extend it later.

It’s WashedBlue: https://iczelia.net/about/#my-setup

Her blog is worth a read: https://iczelia.net/

The problem with forks like that is that the authors later realise the burden of having to maintain a complex piece of software and then they fizzle out.