If I/O speed is important the challenge will be getting lots of nvme slots in a small form factor. Many atx motherboards have bifurcated pci-e slots that can be converted to manage 2 nvme drives at once (in addition to on-board nvme slots) but I don’t know if matx boards do that, so if you wanted 3+ drives that would be the first thing to consider. If you just want a bunch of sata ssds there are more options, but all considerably slower.
Yep, I didn’t know if those were reliable tho or if YouTube had started inserting algorithmic content into playlists yet. And critically, when a playlist ends it should end, not roll right in to other “related” content. And of you’re it’s YouTube-only, so there would be no way to l a playlist that contained other sources.
I’m only hearing upsides here.
Yes, I was thinking something YouTube (or long-form media) specific with the purpose of replacing the YouTube algorithm. So for example maybe someone who loves astronomy puts together a playlist about supernovas that include a video from pbs spacetime, Sabine Hossenfelder and kursgesagt. Then he puts together other playlists. Over time you’d recognize this curator’s name and trust them not to point you to garbage.
Maybe it’s time for the return of curated lists. Back in the day (even before the web) there used to be people who organized lists of videos to rent, books to read, etc. why not bring it back in a lemmy channel or some other trusted service. Just curated lists of specific videos.
Do it! Do it! Do it!
😂
No, it’s better than that. It’s a chromium browser that does the browsing for you!
The front page of the federated web 🙂
If it’s any consolation, at least a kilo of coffee is many more servings than a kilo of beef.
I buy DVD/Blu-Ray collections and rip them. Facebook marketplace (the only thing I would ever think of using FB for, and obviously in a containered browser, etc), OfferUp, EBay, etc make it more affordable, and also really make me realize how little I want to bother having a copy of.
One can dream…
I’ve you ever looked at CUPS code? Forget wizardry, it’s straight up Eldritch lore in typed form.
So yesterday’s Anthropic victory said that the use of copyrighted works to train LLMs was fair use, but only because Anthropic had been buying and scanning in the media so the first sale doctrine was protecting them. Did the judge in the facebook case now really say that it was ok to pirate the materials in the first place because there was no market harm? That seems impossible, it seems like it would open up a huge loophole in copyright law.
Ah the old “owe $100 and the bank owns you; owe $100,000,000 and you own the bank” defense.
FYI many credit cards will allow you to make virtual numbers which you can then use for individual purchases or groups of purchases. They were originally meant to help combat fraud but I use them for subscription services so that if they give me a hard time about quitting I can simply cancel the card number and file a dispute (which in my few experiences has always gotten me my “overage” fees back)
The safest and most flexible option would be to configure the BIOS to not use the RAID controller and just “see” the drives as a regular JBOD, and then setup a ZFS RAID-Z1 array and configure your zvols, etc. in that instead. RAID-Z1 is functionally very similar to old-school RAID-5. There’s virtually no performance penalty with software RAID these days, and you’re eliminating the proprietary RAID controller as a single point of failure.
He’s a modern day arms dealer.
On the one hand, I understand the inherent limitations of pseudonymous social media and why a corporation and even end users might benefit from authoritative user identification.
On the other hand, oh hell no.
This is only really true if there will never be a need to really learn the code base — which maybe will be true for most devs, it’s hard to tell. But if not it could leave the majority of juniors with a much shallower understanding of the way things work under the hood.