Even if it’s not true, it’s not surprising.

This is a weird sentence. I honestly have no idea what it means.

Oh, forgot to mention: striping in ZFS will use the capacity of the smallest drive. It sounds like you have a 1TB drive and a 4TB drive, so striping would give you access to 2TB at most.

Losing one drive in a striped pool with no redundancy means the entire pool is shot. Restoration from your HDDs may take a very long time, on top of data loss between the time of failure and your last snapshot. Striping without redundancy is fast, but dangerous.

This may work at first, and maybe you really do have a use case where this kind of failure is tolerable. However, in my experience, data is precious more often than it isn’t. Over time, you’re more likely to find use cases where the loss of the pool will be frustrating at best, and devastating at worst.

If you’re not using any redundancy, I would create separate pools so each drive can fail independently. You’ll have all 5TB of storage, but not contiguously. That at least constrains the failure modes you’re likely to run into.

If you are striping with redundancy (e.g., RAID-Z1), which I would highly recommend, you can lose a drive and not lose any data. That would take at least 3 equally-sized drives though, and you’d only be able to use the capacity of 2 of them.

I agree with you, however Jellyfin is not intrinsically more secure than any other piece of software. You have to be very careful how you go about deploying it if you open up external access, as you are dependent on the Jellyfin devs to fix vulnerabilities and they aren’t actually being paid to do this. If you’re paranoid about privacy, you should be paranoid about this too; the people sending subpoenas aren’t above port-scans on ISP subscribers, they did it back in the early days of torrents.

You get control and privacy, but you also get responsibility. It’s a trade-off, and one I’d certainly make if Jellyfin were more mature. That’s just me though, I’ve been hosting my own stuff for about a decade now and I can set up an isolated environment for Jellyfin to run within. Plex is a lot more newbie-friendly and I’d still recommend it for most folks unless they for sure know what they’re doing.

As an aside, these concerns are common to all FOSS software that don’t have deep-pocketed backers. Jellyfin is likely never getting those, unfortunately. I hope they can find some other way of sustaining themselves, they’ve not got much money for the scale of development needed and it’s all volunteer-driven today.

https://opencollective.com/jellyfin

I want them to keep going, and I’ve even donated to them. I still don’t think it’s at a place to replace Plex for most people yet though.

Left a comment further down: https://lemmy.world/comment/13599910

Switching between wasn’t seamless, it kept forgetting where I left off on the last device, which was pretty annoying. Also, mobile/remote connectivity was spotty for me. Never got to the bottom of that, but my best guess is Plex’s relay system makes up for a lot of random network issues. My best work-around was to add my phone to tailscale, but obviously that’s not a great solution and won’t work for a lot of devices.

Overall, my impression was that Plex is a lot more polished. I also bought a lifetime membership years ago, so I have no incentive to switch to something that isn’t better. Plex isn’t perfect, but it was still better than Jellyfin as of a few months ago. I honestly hope that changes soon, I have zero faith in Plex as a company.

I’m in the same boat as you. I’d love to switch but the user experience of Jellyfin is still pretty bad outside the most basic cases. If you have a media center PC, it’s fine, but if you want to be able to switch between several devices the way you can with Netflix, it’s quite poor.

Plex is slowly trending down and Jellyfin is slowly trending up. I hope Jellyfin outpaces Plex before the enshittification is complete, but it’s a steep hill to climb.

The math here is beyond me, but this statement from the paper seems contradictory:

The obtained equation is covariant in space–time and invariant with respect to any Planck scale. Therefore, the constants of the universe can be reduced to only two quantities: Planck length and Planck time.

Planck time is derived from the speed of light and the gravitational constant. So wouldn’t there be at least four universal constants?

It’s really more of a proxy setup that I’m looking for. With thunderbird, you can get what I’m describing for a single client. But if I want to have access to those emails from several clients, there needs to be a shared server to access.

docker-mbsync might be a component I could use, but doesn’t sound like there’s a ready-made solution for this today.

Yeah, they are ideally the same mailbox. I’d like a similar experience to Gmail, but with all the emails rehomed to my server.

Steam + Proton works for most games, but there are still rough edges that you need to be prepared to deal with. In my experience, it’s typically older titles and games that use anti-cheat that have the most trouble. Most of the time it just works, I even ran the Battle.net installer as an external Steam game with Proton enabled and was able to play Blizzard titles right away.

The biggest gap IMO is VR. If you have a VR headset that you use on your desktop and it’s important to you, stay on Windows. There is no realistic solution for VR integration in Linux yet. There are ways that you can kinda get something to work with ALVR, but it’s incredibly janky and no dev will support it. There are rumors Steam Link is being ported to Linux, nothing official yet though.

On balance, I’m incredibly happy with Mint since I switched last year. However, I do a decent amount of personal software development, and I’ve used Linux for 2 decades as a professional developer. I wouldn’t say the average Windows gamer would be happy dealing with the rough spots quite yet, but it’s like 95% of the way there these days. Linux has really grown up a lot in the last few years.

I do quite like the stability of Cinnamon/Debian, and I think this problem is solvable (even if I have to solve it myself). I generally do not want to spend a lot of time futzing around with my desktop environment, but this is one thing I need to have.

I saw that and tried it pretty early on. That just moves the screen, it doesn’t fill the quadrant.

Updated to be specific, I’m using Cinnamon. Muffin is the builtin tiling window manager for Cinnamon and it does exactly what you’re describing. The problem is that it moves tiles, it doesn’t absolutely position them. You have to keep moving tiles around to get them where you want them, Rectangle just has hotkeys to immediately place and resize to fit the active window for each quadrant that it supports:

• ctrl+cmd+left: top left quadrant
• ctrl+cmd+right: top left quadrant
• shift+ctrl+cmd+left: bottom left quadrant
• shift+ctrl+cmd+right: bottom left quadrant
• alt+cmd+left: left half
• alt+cmd+right: right half
• alt+cmd+up: top half
• alt+cmd+left: bottom half
• alt+cmd+f: full screen

It’s hard to express how natural that feels after using it for a bit, and I’m still using a Macbook for work so the muscle memory is not going away.

I didn’t say it wasn’t amazing nor that it couldn’t be a component in a larger solution but I don’t think LLMs work like our brains and I think the current trend of more tokens/parameters/training LLMs is a dead-end. They’re simulating the language area of human brains, sure, but there’s no reasoning or understanding in an LLM.

In most cases, the responses from well-trained models are great, but you can pretty easily see the cracks when you spend extended time with them on a topic. You’ll start to get oddly inconsistent answers the longer the conversation goes and the more branches you take. The best fit line (it’s a crude metaphor, but I don’t think it’s wrong) starts fitting less and less well until the conversation completely falls apart. That’s generally called “hallucination” but I’m not a fan of that because it implies a lot about the model that isn’t really true. Y

You may have already read this, but if you haven’t: Steven Wolfram wrote a great overview of how GPT works that isn’t too technical. There’s also a great sci-fi novel from 2006 called Blindsight that explores the way facsimiles of intelligence can be had without consciousness or even understanding and I’ve found it to be a really interesting way to think about LLMs.

It’s possible to build a really good Chinese room that can pass the Turing test, and I think LLMs are exactly that. More tokens/parameters/training aren’t going to change that, they’ll just make them better Chinese rooms.

Maybe this comment will age poorly, but I think AGI is a long way off. LLMs are a dead-end, IMO. They are easy to improve with the tech we have today and they can be very useful, so there’s a ton of hype around them. They’re also easy to build tools around, so everyone in tech is trying to get their piece of AI now.

However, LLMs are chat interfaces to searching a large dataset, and that’s about it. Even the image generators are doing this, the dataset just happens to be visual. All of the results you get from a prompt are just queries into that data, even when you get a result that makes it seem intelligent. The model is finding a best-fit response based on billions of parameters, like a hyperdimensional regression analysis. In other words, it’s pattern-matching.

A lot of people will say that’s intelligence, but it’s different; the LLM isn’t capable of understanding anything new, it can only generate a response from something in its training set. More parameters, better training, and larger context windows just refine the search results, they don’t make the LLM smarter.

AGI needs something new, we aren’t going to get there with any of the approaches used today. RemindMe! 5 years to see if this aged like wine or milk.