b3nsn0w

There was a man at CERN once who was sick of questions. His name was Tim-Berners Lee.

i mean, good point on the project size, but buying a domain is honestly such a basic thing that it still feels like a weird result on that equation to me. for fmhy.ml, specifically, i understand their choice, since pirate sites do tend to be quite nomadic with their domains, and the fediverse being so domain-specific is a new thing.

still, domains are hella cheap. the $10/yr figure i quoted is for a “serious” one on one of the major tlds, you can get away with much less if you’re willing to go for a somewhat more niche but still reputable one. especially one that’s longer than two characters.

lmfao, frickin seriously? you’re gonna build up an instance where the domain is part of all of your users’ identities and you’re not even gonna spend the $10/yr to keep that solid? with how much time goes into running a lemmy instance and not getting overrun by bots, that’s an absolutely ridiculous assignment of resources

shameless self-plug but i host an instance for the life is strange fandom (pricefield.org)

i’ve also seen a game preservation instance at preserve.games

which B is hard?

Only reason I see is because of phones breaking. My current Mi 10T Lite was great for the first two years, then it started getting annoying. I can no longer use Wallpaper Engine because of a stupid system update, notifications started getting stuck, sometimes it has other minor annoyances. The hardware is still fine, there’s no reason this phone shouldn’t work, but it doesn’t. Xiaomi clearly wants me to go buy another phone.

So I did. Just not from them. My Fairphone should be arriving any day now. My friend already got hers, and she got me super excited for it.

hey! i’ll have you know i’m only 26. calling me out…

i also haven’t used linux for a while but i’m currently procrastinating on setting it up on my laptop because windows modern standby hella sucks

We’re still waiting for actual figures, the 7% often cited was from weeks before the protest. Reddit Inc’s behavior was extremely unprofessional, they did an absolutely terrible job at controlling their community, and the quality content that’s the main driver for Reddit’s success has took a massive hit as they alienated their core niche. They will undoubtedly vilify said core niche in their communications, in an attempt to fool prospective investors into thinking they just “got rid of a bunch of nerds who were against their totally sane monetization practices”, but what they really did is they cut off the 1 from the 1-9-90 rule, they drove away their core contributors who kept the other 9% of users engaged and the remaining 90% lurking and still consuming ads.

The real impact to Reddit’s platform isn’t going to happen today or tomorrow, the damage will take months to set in as the reduced value of content results in reduced user engagement and retention. And by the time they see the charts, it will be too late to act. As for their valuation, it is up in the air due to the delayed effect, but any smart investor should see the clear signs shown here and exercise extreme caution about the valuation they assign to Reddit. The site appears to perform better for now, because it’s easy to force dissenting voices off the platform, reign in unruly mod teams and force them to open up their communities to their specifications, but pulling the lost users back on the platform, especially their most valuable and connected contributors whose trust and buy-in is now thoroughly broken, is not something Reddit can force. And it’s crystal clear that Reddit can no longer accomplish anything regarding its community without use of force. They lost the carrot and only have a stick now, and sticks don’t bring value to your platform.

besides, talking about news, reddit or not, is one of the main things communities on aggregator sites like lemmy or reddit do

bUt tHe bOtS wIlL hAvE tO pAy FoR tHe ApI tOo

i honestly wonder if they’ll try to sell investors on this absolutely braindead take. after all, easier to ask for forgiveness five years later in court when you already appointed the right scapegoats

Reddit has been careful to set the goalposts entirely in the realm they control, they ensured that in public communication “victory” means having the remaining subs open up. Ultimately, they do have final say over what is actually served on reddit.com. However, what they cannot control is their users, the contributors who built their empire for free. And they did a piss poor job of keeping us around.

They can force mods out, but they won’t be able to force them back in. As for users, I have no doubt they managed to push away the ones most resistant to monetization, but if that really was their strategy, whichever moron came up with that really needs to google the 1-9-90 principle.

i don’t think there’s an api endpoint for that yet. the database definitely stores it, it would be doable, but it’s not implemented yet in the backend

i had mine up for about a week, been running lcs for quite a few days, and both actively use other spaces (like this one) and post a lot in the ones on my instance. currently i’m sitting at 4.5 GB for pictrs (not a typo, that’s what the picture server is called), 2.3 GB for the postgres database, and 5.4 GB for docker. total disk usage is about 14 GB for now, i expect it to grow in the future but idk yet how fast it will be. people are reporting about 100 MB a day since the reddit migration, and tbh that might check out.

if you’re hosting it on your own nas you’re probably gonna be fine space-wise. i’d just recommend to layer a vpn and/or a cache in-between – i don’t know exactly how to do this, i went straight for the cloud route, but i have seen people in !selfhosted@lemmy.world doing that, and the lemmy admin matrix chat is nice too.

just fyi, your instance does have to be reachable on a domain if you want federation to work. also, keep everything you can on the defaults and only change things one by one, the error messages are not very helpful. i spent like a day trying to debug why lemmy wasn’t starting up at first, turns out i just had an instance name longer than 20 characters.

broke: register on one of the main instances
woke: register on a small but well-run instance
bespoke: spin up your own instance

the problem is you’d need to check for visual similarity, not just the same. spammers and scammers of any kind often take advantage of the full power of unicode to create names that look like the same, but aren’t, and even without unicode you have the issues of I vs l, O vs 0, rn vs m, and so on. if we can figure out a check that the impersonators could actually have a hard time evading that would be nice, but otherwise i don’t really see it.

imo we should make the colors non-invasive by just minimizing their visual impact. painting the entire comment with the assigned color would be intrusive, but i see no issue with just coloring the usernames, in a very discord-like fashion.

yeah, the point is that if hyazinthe@feddit.de hashes to, say, blue, they can try to find a similar-looking username that also hashes to blue, therefore helping with the impersonation. if you hash a client nonce that’s different for everyone, you may hash to blue on my screen but green on yours, and there will be no relation between who hashes to which color on your screen or mine. the impersonator will have no way to guess if their name would match colors on either of our screens, and if we have, say, 25, colors, it will be a static 4% chance no matter what they do.

sounds like the self-appointed ruler of some trash heap

display names kinda run counter to this and I’m not certain they’re a good idea

i think they would be a good idea if they worked like they do on mastodon: you get the display name and profile pic displayed prominently, but you still have the full username displayed below, with the domain included.

damn, he was killing off subreddits before it was cool

i think this could be resolved by assigning a color to each user based on a hash. maybe people would try to find collisions there (i.e. specifically find usernames that get the same color as you), but if you do something like color_index = hmac(user_address, client_nonce) % color_count where client_nonce is unique to each client, it would be impossible to manipulate usernames to get a collision or even a higher chance at it.