This is literally one of the best use cases for LLM. I mean hell deepL exist and people aren’t angry at it. Tho it’s not really a general LLM.
But yeah, anyone being upset someone is using a language model for one of the few legitimately useful and good things it provides us. Is just being a hater for the sake of being a hater.
Still don’t consider it a proper back up personally.
It’s just not a good long term solution. A short term backup for redundancy while actively working on something sure. Just keep it local.
But as a real actual backup? No, local doesn’t cut it.
Your welcome and good luck!
May your journey be safe and your disks stay spinning.
Not knocking zfs it’s good at what it does but man I love how culty zfs fans are.
Y’all are dedicated to an extreme for a file system. Lol
Literally everything is vulnerable to supply chain attacks…
Your comment is basically saying don’t go in the sun you will get sunburned so you should always avoid the sun forever.
You HAVE to accept some level of risk if your using literally any software that you don’t make, host, and maintain yourself.
So unless you have actually decent suggestions on ways to prevent issues, improve security or some tips for a new user for best practices.
All your doing is fear mongering and being extremely unhelpful.
I manage it with netbird.
Just have em install netbird, if it’s on they can connect to my website. If it’s off they can’t.
It’s literally three steps for each person, there’s ways to make this easy. It just requires the host to put in the leg work
Slap a vpn on her computer or but a shitty dell from good will. Point the Roku at that and forward the traffic…?
Like this is a solvable setup and forget about it for the next 5 years. Sort of set up.
The only reason to expose it raw to the internet is because your hosting a piracy service.
You can easily keep everything inside of a wireguard VPN and expose nothing directly.
No nix is the new arch.
Lol
I can confirm I had the bright idea to automate audiobooks… Only to quickly realize I suddenly had the problem of how the hell do I organize it, find whats next or make sure I get what I want.
Ended just making a janky little site that’s basically a knock off audible suggestion engine.
Then rigged it up to the automation. Can put things on the list it searches once a day for anything on the list and downloads it if I don’t have it.
It even checks the cover art, tags, and what not to make sure it’s all accurate then slaps it in and sends me a nifty little notice when it’s done!
I’m working to do the same thing with music next.
Shotgun automation is not the way. Curated automatic is!
Tho I do double check everything before first run… Been bit before. Lol
Three models my self, qwen 27b as a dev assistant, a small model rigged up to handle voice input as an assistant that’s connected to my intercom system though out the house. Mostly to run my lights/music/audiobook play back. It’s fun having jarvis control the house after all. Then a image generator.
Llama and swap on a 7900xtx.
I haven’t gotten around to setting up phone access but wireguard for remote access form work in my laptop.
Honestly it’s all basically over priced toys. I’m baffled that people rely on this stuff. I look at it the same way people do with model airplanes or RC cars.
Over priced toys that are hell of a lot of fun. Break frequently and require more care then they are probably worth to use safely lol
I’ve been eyeing lemonade as I am on llama and swap my self. How much of a pita was the migration. I’m a lazy bastard.
Zero or one backups depending on the type of data generally.
For a system if it can’t be redeployed from a list of requirements then I Foundationally ducked up. It means either I’m depending on something fragile, unknowable or over complicated.
Raw data on the other hand goes on a raid array. If I need more redundancy then that, then a single extra off site backup on a cd or something.
A back up on the same drive as the original isn’t a backup. It’s a waste of space, and a stupid idea.
A system that requires excessive nonsense is not reliable.
If I have a on and off site back up and both fail. Then either the world is literally fucking on fire and it doesn’t matter. Or someone dropped a bomb on my storage.
Either way I have bigger problems at that point.
The problem tends to be it defaults to the old fashion way of doing thing. Since most training data is that old fashion approach
If you just remember to prompt it to keep things simple or ask it for the easier way after the first step it reduces that problem to effectively zero.
It’s no different then asking some greybeard how to do something then bitching that it’s too complicated.
They always default to “their way”
Ollama I find is just a waste of existence. It endlessly seems worse then any other option
There’s decades of information to shift though and AI knows just enough to shift though the chaff.
Its honestly likely going to slowly become one of the few real use cases for the tech for the avg person. But from a local perspective. Needing a centralized search engine is likely going to become iffy and lower quality even if they shove ai into it.
I’ve always been confused by the killing power argument.
A sufficiently high velocity impact to the correct location kills regardless of the material in question typically.
So as long as the material does not cause fov to go to absolute shit or something then functionally it shouldn’t matter.
The landscape changes frequently even sometimes two of three times a week.
I use whisper and various whisper variants generally. But I’ve recently tried hy-mt2 and found it to be really good as well.
Hy-mt2 is a tencent model so it may do Chinese really well. My personal use case is translating Bollywood and Chinese movies to English.