NewOldGuard

@NewOldGuard@lemmy.ml

• ⁨Comment⁩ on ⁨From F-Droid to emulators, here's who's hit hardest by Android's new verification rules⁩ ⁨⁨1⁩ ⁨day⁩ ago⁩:

  “Side load” is just language used to mean installing software outside of the official approved vendor store. Linux lets you install software from wherever you want, you’re not locked into a vendor or manufacturer ecosystem. So in a sense all you do is “sideload,” it’s just normal non-corporate OSes call that “installation”

• ⁨Comment⁩ on ⁨The US bans all new foreign-made network routers⁩ ⁨⁨3⁩ ⁨days⁩ ago⁩:

  I’ve got a GLiNet router with OpenWRT, running adguard on it. Best router experience I’ve ever had. I wonder how quick this ruling takes effect, might be smart to buy another while I can lol

• ⁨Comment⁩ on ⁨Do you stick to the same linux distro across your devices?⁩ ⁨⁨2⁩ ⁨weeks⁩ ago⁩:

  I used to use a variety. I’d use Arch on my desktop/gaming machine, Fedora on my laptop, and Debian on my server. But I got the NixOS bug a few years back and now I use that everywhere. It’s great to have every change and configuration documented and available for easy review or modification, and built in generation rollbacks are a lifesaver.

  Thinking of building an HTPC from some spare parts, and I think that’ll be the machine to buck the trend. Bazzite will be everything I need out of the box for that purpose without any effort for maintenance. It’s not getting customized or doing anything but games and media

• ⁨Comment⁩ on ⁨How to reach different services via name instead of ip?⁩ ⁨⁨4⁩ ⁨weeks⁩ ago⁩:

  Can also use mullvad’s DNS, it’s free

• ⁨Comment⁩ on ⁨I wrote a blog post on selfhostesd software to be more organzed⁩ ⁨⁨5⁩ ⁨weeks⁩ ago⁩:

  Great post, also I love your about page 🚩✊

• ⁨Comment⁩ on ⁨Western Digital details 14-platter 3.5-inch HAMR HDD designs with 140 TB and beyond⁩ ⁨⁨1⁩ ⁨month⁩ ago⁩:

  Yalla

• ⁨Comment⁩ on ⁨Would a fediverse alternative of Discord be possible?⁩ ⁨⁨2⁩ ⁨months⁩ ago⁩:

  I exclusively use Mateo for social experiences. I’m in servers and group chats for clubs, communities, and even Lemmy instances. Works great

• ⁨Comment⁩ on ⁨⁩ ⁨⁨3⁩ ⁨months⁩ ago⁩:

  I’m the type of fish to brag about eating the bait

• ⁨Comment⁩ on ⁨Have you know???.⁩ ⁨⁨6⁩ ⁨months⁩ ago⁩:

  It depends on the model but I’ve seen image generators range from 8.6 wH per image to over 100 wH per image. Parameter count and quantization make a huge difference there. Regardless, even at 10 wH per image that’s not nothing, especially given that most ML image generation workflows involve batch generation of 9 or 10 images. It’s several orders of magnitude less energy intensive than training and fine tuning, but it is not nothing by any means.

• ⁨Comment⁩ on ⁨Have you know???.⁩ ⁨⁨6⁩ ⁨months⁩ ago⁩:

  The training is a huge power sink, but so is inference (I.e. generating the images). You are absolutely spinning up a bunch of silicon that’s sucking back hundreds of watts with each image that’s output, on top of the impacts of training the model.