There’s a wonderful android app called Imagepipe. Does a few things, but stripping exif data is one of them. Workflow is also great!

I realise that won’t help against people simply recognising your cat, but it’s useful for protection against bots and stuff.

Tell me you don’t understand what you’re talking about without saying you don’t understand what you’re talking about.

Filled in the survey. A few notes:

• Some of my answers make no sense on the surface - like the “experiment with new technology” block (4 questions). I’ve answered “Agree” to all of them, because I have taken time into account, which is not represented on the questions. Long story short - I do love experimenting with new tech, I’m almost always the first one to try something among my peers, but at the same I never blindly jump in (I’m hesitant) as most of the “new technology” is just
  • Someone repackaging foss and relabeling it
  • Some LLM bullshit
  • An inferior product to what already exists

There are also scenarios where I have already found something that’s the best solution for my case, so I won’t even bother looking at something new, even if it might be the best thing since sliced bread for someone else.

• TIme and effort setting up/maintaining (4 questions). It doesn’t take much time nor effort to set anything up now, but it did when I was starting out initially. I knew very little and a bunch of concepts hadn’t clicked, yet, so it took me days to set up Nextcloud and about half a year (on and off. Probably a week or so if it were all squeezed together) for email.

• The performance and intent to use in the future questions are weird - they feel like the same question, just leveling off in intensity. I’ve selected the same answer for all of them. They probably should’ve been a single question with agree/disagree options swapped for intensity levels.

Good luck with your PhD!

What’s this bear-porridge symbiosis I’ve been seeing lately?

Is this from some sort of q&a?

That is not at all the case in this context.

Mostly agree. Audiobooks are not my thing, but of it were - I’d look for a way to resume where I left off, maybe some recommendation on what to listen to next.

In general - once you’re into hosting stuff and past the initial barrier of setting everything up - adding another service is dead simple.

Can I be unreasonable? I’m gonna be unreasonable.

Gentoo.

.dev domains are required to only be reachable via https. You’ve not mentioned that in the post, so I’m guessing port 443 is not serving or even listening.

I’d delete the screenshot with your IP visible. You never know…

But he doesn’t even wield absolute power. The bit he does hold seems to have been enough.

Majority of openrc/hardened/selinux binhost setup is done, need to figure out the small things.

Lemmy was also giving a bit of a headache, fiddled with limits some more.

I’m fairly certain there’s been an attempt to play with some opnsense config, but there was only time to install the updates. Or maybe this was last week 🤔

Whole path has to be accessible, not just the file itself. All dirs above the file need to have the executable bit set that affects the user accessing the file.

Lemmy.cafe checking in

So does mumble!

I’ve not read all of it, but if you’re referring to the stuff at the beginning - none of those limitations apply to 5700xt.

If you mean something else - then, naturally, I would ask if it actually affects your media in the first place. It might, but I wouldn’t expect that.

I mean if you’re keeping the GPU - you can just set jellyfin to use VAAPI and utilise the gpu that way.

• https://trac.ffmpeg.org/wiki/Hardware/VAAPI
• https://wiki.gentoo.org/wiki/VAAPI
• https://wiki.archlinux.org/title/Hardware_video_acceleration#Configuring_VA-API

in case you want to tell me what I have is fine and I don’t need an upgrade

What you have is fine and you don’t need an upgrade 😁

But we’re not looking for fine, are we? :)

I would keep the gpu and get as many cpu cores and ram as my budget allows. Once you cross into “stupid amount of RAM” territory you can start utilising tmpfs for transient things such as jellyfin transcode directory to:

• preserve those precious ssd writes (not really relevant anymore)
• make it more efficient (feels-good kind of relevant)
• running a filesystem in ram is really cool (most relevant, naturally :D)

No, comercial IPs are fine. You’ll have trouble with some of them - Digital Ocean is a notorious example - where the provider itself blocks outbound port 25 and there’s nothing you can do. I think DO only does that for new accounts.

I myself am running it on Linode - it did get purchased by Akamai a couple of years ago, so I can no longer blindly recommend it - but so far it’s been working fine. One thing I did recently discover was the ability to request a /56 block on Linode - my pre-assigned IPv6 got blacklisted somewhere as at least the whole /64 and simply generating another IP from the same /64 did not help. Getting a fresh block solved it for me, though, and now I know that if this /56 gets blacklisted - it’s my fault. Unless, of course, I get caught up in a /48… 😳

You won’t be able to host email on a residential IP - all of them are on a permanent blacklist. I understand the money argument - and it’s a real argument - but host your own email is just so cool!

Look, ma, we’re famous!