Right, but I would say the same thing and for a gaming machine, I would much prefer something that did the Arch install for me and worked for most games out of the box.

I’ve found CachyOS to be fairly uncomplicated and it’s gaming tweaks make most things work out of the box through Lutris. I’d probably avoid the standard Arch install for a newbie

This is my setup too. I use WG-Tunnel to manage the VPN connection on my phone. It just monitors whenever you disconnect from your trusted WiFi network and automatically enables the VPN.

Only hiccup I’ve found is wireless Android Auto isn’t a fan of a VPN.

It’s kind of funny that we literally grow straws and just don’t use them. Wheat straw is a biodegradable alternative that doesn’t disintegrate and is often just left in the field after harvest. It might not work for thicker beverages like milkshakes or smoothies, but works great for everything else.

Loblaws (Canadian grocery overlords) sells some.

I just got a 7900 XTX because of the much better performance to price ratio. And 24 GB VRAM should help it last quite a while. The trouble was finding stock in Canada. I imagine this 9000 series will have the same issue.

The 7900 XTX is looking like a pretty good option for a fraction of the price. I think i might go that direction. 24GB also helps future proof it a bit.

Oh yeah, it still gets the job done.

Keeping my eyes open for used ones to upgrade with now that the new series is out though. Gives me an excuse to get the 1080 in my server.

I’m still riding my 1080ti…

deleted by creator

There’s no helping this one.

The internet must be a miserable place for you.

Not that I’m familiar with. I would guess that the limited processing power of a phone would bring a pretty poor experience though.

There are quite a few options for running your own LLM. Ollama makes it fairly easy to run (with a big selection of models - there’s also Hugging Face with even more models to suit various use cases) and OpenWebUI makes it easy to operate.

Some self-hosting experience doesn’t hurt, but it’s pretty straightforward to configure if you follow along with Networkchuck in this video.

I came across ryot a while ago but didn’t have a chance to really play with it. It might be useful for something like this though. Grocy is another that comes to mind.

Recent post from the Proton Reddit account: https://archive.ph/quYyb

Seems it’s not just the CEO’s personal opinion.

I see. That’s good to know, thanks.

Can you elaborate?

Ah, yes that’d probably be a good idea.

I’m not sure I understand. Whatever machine is hosting the Immich server will be doing all the processing. Uploading to there will go as fast as your internet connection/storage speed can go, and ML processing will be based on the processing power of the server. You can also manage the ML tasks if you want (Administration > Jobs and/or Settings).

The first big import will always take time if you have a lot of photos, as will the associated machine learning processing. I’d say you don’t need a GPU like that for normal use though.