realitaetsverlust

Well, you know … “them”

It’s real GPS, but within a sandbox without additional permissions. https://grapheneos.org/usage#sandboxed-google-play

Damn I must’ve overread that. But how isn’t it working? I have a private instance and it has been working without problems apart from connecting to brave, which is rate limiting me immediately. And there’s so many sites to choose as potential providers, losing a few to rate limits honestly doesn’t seem like a big deal.

Graphene doesn’t, they got their own sandboxed google play services compatibility layer. Other OS - no idea tbh.

That’s unfortunate, but not surprising. Luckily, the world works without google.

Get SearxNG and you’ll never need another search engine.

Shit country tbh. You might be able to get a cheap stock android device for your ID and your “actual” device keeps living the good life on a custom rom. Idk how feasible that is for your lifestyle tho.

Yeah, that’s roughly what I thought, in which case switching to graphene (or any other custom rom) would completely “release” you from the proprietary grasp of google and their genius ideas. Because even if they add this 24hr lockdown in the AOSP-Project, graphene could rather easiely revert that change in their fork.

I might have misunderstood that all those years, but custom roms like graphene or lineage all use the AOSP source to build “their” stuff on top. Android itself is all the proprietary shit google adds and probably bunch of proprietary firmware drivers, which isn’t relevant in the slightest if you switch to a supported device.

If your country forces you on a locked down OS for ID requirements, something is wrong entirely. You should absolutely make some noise against that and urge them to allow custom OS. That’s a political issue, not a technical.

I’m pretty sure that custom roms will remove that 24hr wait period from their binaries. Android is still open source and removing a check isn’t going to be very hard for them.

The tech ecosystem is as open as you can. Use linux. Use SearxNG. Use graphene or other custom roms. There’s options out there, you just have to start adopting them instead of just complaining online about having no choice.

Afaik, this is not possible yet. There hasn’t been a way to bypass the security measures yet, unlike with the bug in the tegra chip of the switch, or the exploit for the ps5 (or 4?) that got well known recently. Until such time, what you’re looking to do is borderline impossible.

The code has been generated in it’s entirety with an LLM, as I am not a software developer

What’s the point then? There is countless expense management software. Completely stupid tbh.

Stop vibe coding. It makes you stupid and you don’t learn anything with it.

Because AI bad I guess. I’m not expecting sane behavior on lemmy when it comes to AI or capitalism lmao

Jup. Ollama and OpenWebUI is a great stack to tinker with some LLM models. They’re kinda useful for aggregating large datasets, translations, frontend development and gathering relevant sources for me to read into. Also, Qwen has been amazing in understanding frameworks without documentation and writing one for me. I had to use some self-developed PHP framework for a task once and without qwen, I would’ve taken probably two more weeks to get the task done.

MiniCPM has also been REALLY good at image detection, describing it as accurately as possible, feeding it into qwen who then searches what the object could be and returning the result. I always liked google lense and that stack gave me a TEMU-Version of google lense that isn’t quite as reliable, but definitely very useful.

I’d be very surprised if this was actually as good as written in that article.

I don’t think “selfhosting” and “paid for” goes hand in hand because, at the end of the day, the application somehow will still contact some authentication server or some similar bullshit. That’s the contrary of what most people want from selfhosting.

I think this community should stick to actual OSS, free applications, not some semi-corporate bullshit.

Sorry I’m so green, gotta start somewhere!

Start with the documentation. Docker has a great introductory section that teaches you the basics.

https://docs.docker.com/get-started/introduction/ (the pushing your image part is not that important, the rest is)

Running a project that does things you don’t know is not the best thing to learn. Learning is done by going through the basics first, not immediately firing docker compose, which is one step above pure docker.

I’m using the 35b models.

Quality for qwen is mostly fine - sometimes it does hallucinate some shit while thinking, but it does correct itself almost every time. But the answers itself are, for the most part, precise and useful. Not what you know from the cloud models, obviously, but it’s absolutely fine for everyday use. What is actually annoying is the web search - not sure if that’s a qwen problem or a problem with open webui, but it actually takes a long time to finish the search.

I once had a situation where a model was running into an “infinite loop” while thinking, thinking the same line over and over again. And once, qwen just started outputting chinese halfway through the answer lol.

When it comes to context, I’m gonna be very honest - I don’t know. I have never hit any kind of problems or limits because of that since I’m not using AI over a long term project. I use it for small, concise cases and that’s it.

I use an 6700 XTX and it’s working perfectly fine, depending on the model. Gemma4 takes a long time to generate answers, but the Qwen-Series is quick and starts generating answers in ~10 seconds.