Not a weird example. I have my self hosted video server (PeerTube) and I tinkered with transcription thanks to whisper.cpp locally. It “works” in the sense that most of it is acceptable. It still does mistake though. I provide all my content, including hosting, at my costs and to anyone in the World for free.
So… I definitely see the value. I’m only saying that it has downsides and quality-wise relative to professional, it’s still bad.
Don’t let them know about cp, we’re all doomed! DOOMED! /s
PS: sensationalist BS. Interesting as a philosophical concept for sure but here it’s just fear marketing by AI-bro for their VC-funded scam.
Arguable… it’s OKish at best, definitely nowhere near as good as professional… then IMHO it’s like spotting a spelling mistake in an official document, you instantly look for MORE mistakes then it become distracting. There is something powerful about trust that once it’s broken, it’s hard to get back. Once a spelling or here transcription mistake happens, then we brace for more (rationally so) and it becomes a very taxing endeavor.
So… sure STT progressed quite a bit but it’s STILL not good enough in a lot of cases.
Case in point, IMHO when there is a choice, most people (everybody?) would rather have human made captions than AI ones.
Pretty good for spam and scam.
Nice, I’m replying here using Tridactyl for Firefox which is like Vimium so clearly I’m interested.
Does it work well on KDE (so I imagine KWin here) with X11 on Debian 12 (as it’s not mentioned in the listed distributions)?
Some apps are still done this way, e.g. transmission the BitTorrent client, but also ALL self-hosted Web apps. Sure it might feel a bit much to install containers on your phone “just” for that, or having to go through REST API despite being on the same actual device, but still it provides a TON of app.
Anyway, yes I agree that it is often a better model. Still a lot of apps, e.g. Blender, Inkscape, etc do provide a CLI interface. So one can both use them with a GUI or without. It’s not decoupled like transmission but arguably it covers most needs.
The “struggle” is because Apple and Google refuse to do so as they built the platform to give themselves priority.
One can trivially do so on a Linux phone, e.g. PinePhone with PostMarketOS.
Source: I did it. Plenty of others do through the usual ways, e.g. pipe in the console but also with things like https://sxmo.org/docs/user/sxmo.7.html#HOOKS
I also like ANBERNIC RG28XX and PinePhone/PinePhone Pro with or without the keyboard/battery extension.
My biggest concern is that I have a bunch of games installed on various drives that are all Windows (NTFS?) formatted and I’m not sure if I’ll be able to run them on Linux.
I’d check https://www.protondb.com/ and if they your favorite ones are native/gold/platinum then… move on. I’d initially NOT erase my drives and, assuming you have either patience and/or a fast connection, just let it re-download and install overnight, then enjoy. If need be bring the saves back (but again via Steam, should just work) and only once that’s done, erase the Windows partitions. This is a no risk process. Honestly some games will not work but IMHO this isn’t the question. The question rather is… will you have more playable games than time left, if so, then considering moving even without 100% coverage.
FWIW I’m been using PmWiki on https://fabien.benetou.fr/ (and other of my wikis) for a decade now and it covers most of your need (no DB, track change via wiki page history, link and back-links, tons of recipes to extend) but also has some down sides (PHP, even though improved a lot since early 2000, isn’t exactly popular at the moment, own syntax that’s close to markdown yet isn’t it, not the most popular wiki engine and relatively complex). I’m also using my Gitea instances https://git.benetou.fr/
Honestly though I’m not really sharing this as to advocate for PmWiki and/or Gitea but rather to highlight, and even think as I share, that I don’t have a single tool for both. Sometimes I need notes, sometimes I need code repository, sometimes I “just” keep files in my ~/Prototypes directories. I’d love to have a “unified” solution that manages both notes and code but so far it seems each specialized tool, including issues on repository, does a better job. Maybe it’s just the habit. Anyway it brings me to the question to you :
what do you actually expect the unification of those tools would bring?
Edit: I also use NextCloud to edit documents I need across devices or share with others. Same here, I could “just” upload such files (e.g. photos, rich text, spreadsheets) to my wiki as attachment and yet somehow I use this dedicated tool.
Cool tech but I don’t get it. Who cares for ghostly 3D re-creation of moments? It demands so much more than snapping a 2D photo for a result that can be qualified “strange” at best.
I find XR much more interesting for things that are otherwise impossible, say traveling through the solar system or the human body, playing a rhythm game while punching things in the air, etc.
This is totally overblown. They are not “worlds”, it’s usually 10x10m spaces at most, nor photorealistic, so much is left out, there is not animation, no physics, etc.
PS: FWIW I tried some splats in XR, and I also did some photogrametry few years ago. Again it’s an interesting process but it demands a lot for a result that few non-tech people would genuinely be impressed with to the point of replacing their holidays photos with.
Drivers are 535 on stable, cf https://wiki.debian.org/NvidiaGraphicsDrivers so it’s definitely not the very latest yet even though drivers are important I doubt (and please share benchmarks if I’m way off) there is a radical performance difference.
A base Debian would (if it even works) probably less FPS than an current gaming distro.
Can you please elaborate on why? I’m running Debian stable with NVIDIA drivers and… it “just works”. I’m using Steam to get Proton and game content (e.g. was playing Elden Ring minutes ago). I didn’t tinker much so wondering what I could be missing.
FWIW did you try https://download.nvidia.com/XFree86/Linux-x86_64/435.17/README/powermanagement.html ? Namely enabling nvidia-suspend.service/nvidia-hibernate.service nvidia-resume.service services?
Mind expanding just a bit through? IMHO it’s not orthogonal in the sense that either your backups are :
Isn’t it?
No.
I spend a significant amount of time on other things, e.g. NOT using BigTech, no Facebook, Insta, Google, etc where I would “volunteer” private information for a discount. I do lock the physical door of my house (most of the time, not always) and have a password … but if somebody is eager and skilled enough to break in my home to get my disks, honestly they “deserve” the content.
It’s a bit like if somebody where to break in and stole my stuff at home, my gadgets or jewelry. Of course I do not welcome it, nor help with it hence the lock on the front door or closed windows, but at some point I also don’t have cameras, alarms, etc. Honestly I don’t think I have enough stuff worth risking breaking in for, both physical and digital. The “stuff” I mostly cherish is relationship with people, skills I learned, arguably stuff I built through those skills … but even that can be built again. So in truth I don’t care much.
I’d argue security is always a compromise, a trade of between convenience and access. Once you have few things in place, e.g. password, 2nd step auth, physical token e.g. YubiKeyBio, the rest becomes marginally “safer” for significant more hassle.
Recent 535.216.01 seems to improve that.
Whatever machine is hosting the Immich server will be doing all the processing
By default yes but it’s possible to do remote ML, cf https://immich.app/docs/guides/remote-machine-learning so my “ideal” setup would be all on RPi5 BUT initial ML tasks (because it’s going to be > 100GB of photos at first) either directly on my PC with GPU or remotely (cf guide).
Indeed, professionals are expensive, and IMHO rightfully so. I’m only trying to highlight the fact that it’s unfair to imply it’s much cheaper when the quality isn’t on par.