I just deployed tinyauth and it actually supports ldap and it can read also ldap groups. Gonna tinker with it more tomorrow, but so far I like the simplicity a lot.

Thank you, that is not as bad as I expected. 👍

Unless your brain is hallucinating this whole reality 😀

How much electricity does it burn? And how much does it cost to let it run overnight? I feel like this is the same problem that the companies have, its way too expensive to use it on a lots of things and it is way cheaper to use it on someone else’s server…

I tried, but didn’t want to invest into it, so I used my 12-GB-vram gpu but no matter what I did it was worse than Lumo from proton,which is basically only one LLM from 3rd party that I am okay to use on occasions and since it has really good free tier I gave up on self hosted LLM.

Which country are you from if I may ask? 😀 I think in EU this would not fly. I think at most they can terminate your service if you torrent a lot but even then I can’t say I saw this happen. Usually they start throttling your internet temporarily and send you email that you should not do it.

Can an ISP forbid you having your own home server? That doesn’t sound legal. And if I am not mistaken the cloudflare tunnel is a lot like tailscale (which i would recommend over cloudflare anyway) it doesn’t need open ports or anything like that.

Its per hour. Basically my nas draws on average 40 W at any given moment. And it’s nas but also run apps.

Yes it is true. AI gets better if you know the right questions to ask. Inexperienced people ask the wrong questions because they don’t know the pitfalls.

It’s such a bad experience to have to read ai written emails. I would rather read short grammatically incorrect email than formally well written novel from llm. Write your email, ask ai to fix grammar or check formal vs informal phrases and send it, you will do a favor to your recipient and on top of that you will learn something.

Yeah, the text reads horrible, I had a feeling he wanted to write something profound, but wrote a total slop instead (there is an irony in there)…

I think that is the whole issue. AI should be used as a tool to enhance our skill, instead people are using it as self sufficient entities, the result of which are burnt tokens and unmaintainable and buggy code. Same is valid for other areas. Instead of letting it generate an email - a wall of text that has to be interpreted by another llm - maybe use it to check on your grammar or suggest changes in the text you write.

My nas takes about 40 Wh. And my network machine (proxmox + openwrt) about 20 Wh. When I upgraded my nas a year ago it was actually drawing twice as much, I have since then switched to more efficient processor (amd from normal to E) and I took out graphic card that I didn’t use. I could have in theory merge those two, but I wanted to separate networking from nas+apps. I don’t know how much my smart devices take but I don’t have many so it can be up to 50 Wh total maybe. But that is more smart home than self hosting 😀

Checkout Lumo from Proton. Its quite good and private. You can read about how it works on their blog. Otherwise selfhost it.

I stopped reading at Carl Dong 😂 … No, I read it all, thanks for good info!

I will never trust something genetated by llm and barely reviewed by human. Ai agents already showed they are willing to cheat, lie and then cover it up. No no no.

My number one recommendation is Fedora. Extremely stable, popular (so there are lots of q&a online), easy on resources. If you go Debian please don’t go Ubuntu, its buggy for some people (me included) and snaps are 💩. If you want Debian go Linux Mint, which is very beginner friendly.

The thing about automated encrypted drive unlock is that you are missing kernel check pcr (8 or 9 I don’t remember) and without that anyone can boot with compromised kernel and unlock your drive. Which makes encryption kind of pointless. The same pcr, however, means that you have to rebind after every kernel update (which is quite often on many distros). The disadvantage is that with current state of the software handling auto unlock on Linux is kinda flaky and rebinding may involve more than one restart. I eventually realized that it’s less trouble to just skip this altogether and enter the password every time and then set autologin.

The permissions have to be justified when uploading to flathub (via PR) and the exceptions are not always accepted. Flatpaks ARE safer than system packages and if you think otherwise I recommend you taking a look who can access all passwords on the keyring and who doesn’t. (Getting such exception on flarhub is very hard). Flatpak 2.0 hopefully gets user approvals for permissions. But still with current flatpak you have the option to reject permission via flatseal.

I would remove Brax Phone altogether, that guy is fearmongering people into buying old insecure phones. There were times I was watching his videos until I saw the pattern…

Tbh depending on your threat model you might be ok with just a regular os, flatpak sandboxing and an occasional vm. For example I have VMs for my dev projects, or for tools that I use seldomly but doesn’t trust like android tools. If I need I pass usb in (like if I want to flash phone with the android tools…) You will be much less limited but still have decent protection and those vms start fast. I have Fedora host with Fedora VMs and they start in like 10 seconds.