This is similar to what I do.
I have a USB drive with the whole bootloader + decryption keyfiles on it. I remove it while it is running as everything is stored in RAM and already booted.
Downside being it has to be plugged in to update the boot partition during an upgrade.
And you are often paying 140-200 for a pi nowadays to make it have the same usability as a laptop (pi, power supply, sata hat, data drive because SD cards simply fail after a while under server IO) while you can get cheap used laptops for 0-100.
So unless you are running it for more than half a decade (which rarely happens with selfhosters for a main server), you are probably spending more in total on the pi.
My Lemmy instance admin apparently is also a home assistant contributor. Cool!
Yep, you can get an m.2 NVMe to USB3 converter very cheap and stick any m.2 nvme drive in it. (Also sata versions exist for m.2 sata)
Much safer solution for your data.
Is this the same Bose company that makes the standard Bluetooth ANC wireless headphones that everyone and their mother uses if they aren’t using earbuds?
I had the same thing on Bazzite just with the local network, not a VPN.
I believe it has to do with the firewall. You have to open the port both incoming and outgoing for 53317.
But you literally have to be on the same network, so for example if both devices are on the same local network (hence local in the name) and your phone is on a VPN but your computer is not on a VPN, then it won’t work.
It should work if you VPN into your local network remotely so that both devices are on the same LAN, however, then that won’t work anyway because you have to have physical access to the device to accept the transfer (you could probably use a remote desktop to do that, but then it is getting complicated)
LocalSend.
No more USBs ever (outside of install media). So so simple, fast, and works on all devices and FOSS.
It is really the best UX of any file sharing app I have experienced (outside of airdrop I guess, but obvious problems there)
Okular is also a favorite of mine.
There is also leantime.io that I have been hosting for 5 years or so. It is a bit more than planka or tarallo as far as scope I think, but it has integrated kanban, gannt charts, and hour logging which is all I need for my personal projects.
Your cloud example is exactly right and exactly what we want to NOT HAPPEN.
They shoved the cloud so much down our throats so that they can force you into monthly income-sucking unneeded subscriptions. That is it. That is the single reason everyone did it.
The result is now the average user has a much worse experience overall. One literally has to fight with Microsoft products to save things on their own computer. IoT and smart products literally won’t function without connections to their “cloud”. Phones come without SD card compatibility and with low flash memory to force you into cloud subscriptions. Now every damn piece of software is a way overpriced subscription that almost all originally started as “switching to cloud infrastructure” (fucking adobe creative cloud).
The “cloud” has had so many data breaches and people data have been stolen, siphoned off, lost due to bugs, and sold to earn even more cash on the side.
A huge portion of the general corporatization and bad enshittification of digital services and software in general can be attributed to “the cloud shoving down our throats” that you describe.
AI is looking to do the same thing except castrate peoples’ digital skills, critical thinking skills, transcription skills, and writing skills in order to siphon more and more of your income off in the form of AI subscriptions while they double dip and sell everything you ever say to it and triple dip in mining everything you say to it as R&D that you pay to do
Companies need to do the fucking R&D themselves with their revenue of a small country and stop forcing regular people to pay to be their alpha and beta testers and focus groups, and people gobble that boot up so hard because LLMs have a few small areas where they are slightly useful and can save 10 minutes per day and make them not have to critically think, so people will literally sell their data, their already small income, and their soul to save 10 minutes, and in 10 years the digital experience will be even more shitty and degraded than it got after “the cloud.”
Your usecase is the exact definition as using LLMs as accessibility and to actually better the user experience for certain people which is not the goal of any AI company or 99% of LLM integrations
TD;DR
Non-consentual cloud shoving has caused newer generations to think that paying corporations every month to save files is normal and that your data is not yours and always corporate property ™®©, along with the decimation of understanding simple file structures. You can actually talk to teachers and professors and they unanimously say that tech literacy has nosedived.
Now with the LLM shoving, they are trying to force the new generation to have to pay subscriptions to think, write, compose, draw, and get information by stripping them of those skills.
Also the Node 804 is worth looking into with an entire separate chamber for HDDs in order to keep them cool without exposing them to GPU and cpu heat, plus it is a lot shorter instead, sometimes easier to fit places (mATX motherboard only)
Obsidian ticks all of these boxes and syncthing to sync notes is a 5 minute setup.
Plus it stores things in plaintext instead of a database format that vendor locks you in (despite the claim of “no vendor lock in”)
Ooooo yay another half-baked AI shoved into everything whatever possible.
No, ssds have a ton of wear leveling where data is shifted around and not deleted. Deleting data wears out the SSD, so it is held as much as possible with the controller. SSDs are like 10% bigger than advertised just to prolong the life.
Even if you write the whole thing with random data then zeros, it will still have blocks in unaccessible (to normal users) places that contain old data.
Always best to use disk encryption or keep any sensitive data in filesystem encryption like plasma vaults or fscrypt.
Not really self-hosted, but I set up obsidian with syncthing and am going to transfer all of my notes from book stack to it and let bookstack be more organized documentation and obsidian to be a big scattering of notes and tags and such. I tried it with bookstack, but the flow was too much of a barrier for me to use it consistantly
I am doing something similar. I use OIDC for everything possible.
Authelia is quite picky about everything being correctly populated, but if I remember right, the documentation doesn’t do a great job of explaining different variables for someone outside of the security industry (similar with traefik). I found a good tutorial via search that got all of the defaults set up, then playing with the options to my liking and now it is just copy pasting the condiguration per app that I want to enable, generating an key and hashing it.
If you want, I can sanitize my config and share it?
Also code development programs.
Code OSS (VSCode) in distrobox with clangd language server and NCS for embedded development sucks up like 8GB of my 32 just having 6 or so files open once you start building because it keeps builds in RAM or something. It balloons fast at least lol
Especially for jazz albums. Very difficult to find
There was an old fighter pilot game that I used to play on linux back in the day all the time in the early 2000s. I can’t remember the name but it was because my dad’s laptop was very cool to me and ran SUSE, so I played that, super tux, and a few free games because the alternative was a windows 95 machine with a 10 gig upgraded hard drive.
Lego is actually one of the very few companies that isn’t batshit crazy over video game IP. (Real life Lego clones are probably different though).
They even gave a shout out to Manic Miners (a rock raiders fan remake) on the official podcast and haven’t done anything to take it down. I can’t remember if they officially said they won’t do anything also
You can also look at the MKBHD 2024 smartphone camera comparison test with the FP5. I would suggest taking the test yourself if that is still possible.
I would guess that the camera will be comparable. (Everything below if FP5 assuming about the same performance with the FP6)
For me, daylight pics were after all of the pixels but before anything else. I like the more neutral not supremely over-saturated over-sharpened/smoothed pictures that many phones take nowadays.
For me, it was middle of the pack for dimly lit photos.
For the overall ELO with everyone, FP5 was on the mid-lower end (of a comparison of all flagships + pixel A series), but perfectly usable for people who aren’t doing social media as a job.