LLM coding AI is a really useful tool, not enshitification. They're also entirely optional. Way different then shoving it into random products and services without consent.
for/and levels of health science never imaginable.
let's not forget tnt was both a great and awful invention. I know this is different but there is no putting the genie back in the bottle.
ai will be the new normal for as long as there is normal. all y'all saying its stupid or a fad are like the people who said the same at the beginning of the net.
speaking of which, without the net we couldn't have ai. or at least LLMs.
AI will revolutionize health science in ways previously unimaginable.
Let's not forget that TNT was both a remarkable and terrible invention. While this situation is different, there's no turning back now.
AI will become the new normal for as long as there is a "normal." Those dismissing it as stupid or a fad are like the skeptics at the dawn of the internet.
Speaking of which, without the internet, AI—especially large language models—wouldn't exist.
A handful of already obscenely wealthy people will get vastly more wealthy and we'll all be stuck with the shit "products" they've ruined with AI... Totally not a broken system we have.
This is what everyone wanted though, isn’t it? I mean, we were trying to warn people about this shit for a long time- but we’re downvoted into oblivion for merely suggesting that AI was going to fuck shut up.
AI's most successful accomplishment is going to be its ability to parse the astronomical amounts of personal data corporations and governments have been collecting for the last 20 years. Enshittification and the identification, surveillance, and targeting of dissidents.
parse the astronomical amounts of personal data corporations and governments have been collecting for the last 20 years
Always the plan, funnily enough. They didn't necessarily know what they'd need the data for, just knew it was valuable and should be sucked up and saved until the opportunity presented itself.
I've never used an AI image generator, so why start now? lol. Though I did use Google image search to find "train with rocket on it" which probably does use ML under the hood to classify images. So it's not completely devoid of irony lol.
Congrats. Tesseract has always been very impressive.
This post suggests there's a problem with Lemmy/PeerTube federation, but it'll be good to see the videos embedded (even if they have to be brought through manually)
Interesting about the broken Peertube federation. I'm following a few channels (mostly for testing purposes) and I did get new posts to them from PT's side and saw my votes reflected on each end. Haven't gone back and done any comparison's lately, though. That post says it hasn't worked in ~3 months. If it's a Lemmy issue, It's possible that it works for me since my instance is still on 0.18.5? If something changed with Peertube, then I may have just not noticed. Will check on that when I have some time as I'm now curious.
Update: Seems the ones I got were just from the initial fetch.
Update #2: A new PT video came through this morning on one of the channels I follow. So maybe federation does work but is just wonky/inconsistent? Again, I'm still on 0.18.5 so may have something to do with it.
Regardless, yeah, like you said, they'll still embed when brought over manually.
There has been a huge wave of spam accounts recently, and I recognize some of the names on this list. Of course these accounts gets banned quickly.
Getting a fake email address is really easy, so I don't think it makes any difference whatsoever. On fastmail which im using, I can even create fake ones with a click on a button. It's built into the service to have different emails for different sites.
Having closed signups will not stop any spam accounts. It just adds a delay between registration and being able to log in. There is no way to know if it's a spam account or not before they sign in. Of course, some of the names in the list makes it obvious, but they just change to ordinary sounding names then.
I think it's frustrating for new users to have to wait for a random time period to get accepted, specially when it serves no purpose. I'm not going to be any wiser if I approve the registration now or 12 hours later.
Or am I wrong? I haven't been a troll myself but seems like I would just spend 30 mins to make accounts in random instances and then fire off my bot a few hours later or on the day after.
How do you know which accounts are spam accounts before they even started posting?
If the spammers and trolls manage to make you defederate from Lemmy.today, they will pick another instance. And every defederation will continue to make the fediverse smaller and more centralized. Sooner or later, small instances may even stop accepting new users for fear of being defederated from the larger network.
I would like to stop having these spam accounts and trolls on the instance. I could add email verification but it won't make a difference and just make it more annoying for new users. I really don't want to punish them for what trolls and spammers are doing.
But sure, I prefer it to defederation. So if you think it makes a difference, I could add it just as a sign of good faith. But it's really a pointless gesture.
tesseract.dubvee.org
Active