You must log in or register to comment.
Oh no who could’ve possibly seen this coming? Literally everyone, it turns out.
Good read. I love that suburb meme as an analogy of what the internet is going to look like. IMO it’s already there.
Thank you :D I also thought it was a good analogy, especially since we’ve just accepted it as inevitable. Even with all the urbanism revival enthusiasm on the internet, they never push for beauty, just practical stuff like walkability, public transit, etc. It’s good stuff, but I want bread and roses too.
And yes, it’s getting there fast, if it’s not already there. I remember in 2015, when people still loved google, and I started talking about what I then called “The Apple Crisp Problem.” In a span of just a couple years, googling recipes went from super useful to entirely SEO blogs of maybe-not-real women in their late thirties named “Kate” taking their dog named “Pancake” to the orchard to pick the perfect apples for her also-not-real nana’s apple crisp recipe. Recipes were one of the leading indicators. Now it’s just everything. Super lame.
That’s two twitter clones from two tech billionaires in just six months. Capitalism truly breeds innovation.
What’s the other Twitter clone (other than Threads)?
Bluesky!
It’s not “ai”…it’s just a few billionaire assholes.
& it’s not just the internet. It’s the entire “reality”.
deleted by creator
TL;DR; - The internet is getting shittier at ludicrous speed, thanks to AI bots. Expect the next generation of AI to be even worse, as it’s likely to be fed the shit the current AI is dropping everywhere.
Who wants to bet that the powers that be will make a CAPTCHA that you have to decide if the text is written by human or AI.
Reminds me of an iconic film…
oh god, human centipede??
It’d be funny if a lot of bots get discontinued just because the tools can no longer discern real and fake content and just become unusable.
They already can’t. They just rely on the assumption that most of the data they collect is correct. Which is generally true, there is more correct than incorrect content on the internet. The inability of the bots to discern incorrect data coupled with their ability to make it sound authoritative is what makes them dangerous.
What’s interesting is this is also kind of a circle of tech moment too. At least for me, search has been sort of killed (or google search anyway), but it’s just back to the Internet of 1997 again, where we have “sort of useful” “search engines”, some walled gardens like AOL was, and maybe webrings or the original sort of Yahoo! curated link / subject sites / lists.
So be it. Humanity got along just fine without the internet for milleneia. Its clear at this point that the public internet isn’t going to be used to improve our human existence. If there is anything to bemoan, its that the practical usefulness of the internet will only continue to get more and more narrow.
Right but how much of our daily existence is tied to the internet? Like all of our banking systems, our commerce, communications, infrastructure.
I think that functionality will contunue just fine. It’s surfing that dying. They’re trying to put up billboards up wherever are eyes are looking and we can’t even see what we’re looking for anymore. AI is just a “smarter” billboard. Eventually we’re just going to stop looking.
Yup. The Internet used to be the yellow pages. Now it’s turning into an offramp.
deleted by creator
We got a along without antibiotics for millennia, too, but I wouldn’t want to lose access to them.
Speak for yourself. The internet has been a great place imo, at least partly.
Google made the recipy for the shitty internet cake and put it in the oven. AI is the icing.
deleted by creator
Only thing I heard about zoom is their back to office?
They also gave an opt-in to allow all your data to be used to train AI models. You auto opt-in if you use any of their AI assisted services
They changed their TOS to allow themselves to license everyone’s videos for A.I. training (or anything else). One of the execs tried to say they weren’t doing that but unless they change their TOS, they can and no doubt will.
For some people, that’s a personal privacy issue but for people who have Zoom calls about, for instance, health records, it makes Zoom illegal. And even if it’s not illegal, companies use video calls for discussing proprietary information they don’t want to be potentially licensed to competitors.
Does software exists that encrypts video and audio data on one end, and requires a key to decrypt on the other end? Anyone looking at the feed without keys would be seeing garbage.
Zoom already does 256bit AES end-to-end encryption. From what I understand it isn’t the live calls but the files and recorded calls you save on their servers after are what they would use for AI training.
Zoom already updated their TOS a few hours ago to supposedly address the issue. https://gizmodo.com/zoom-ai-privacy-policy-train-on-your-data-1850712655
Update, August 7, 5:06 p.m.: After this story was published, Zoom issued an update to its Terms of Service. The article has been updated to reflect the change.
Zoom and MS Teams both seem to have E2E encryption for 1-on-1 calls. They own the code, tho, so whether they really cannot decrypt the stuff is a matter of trust.
Oh,forgot to mention there is support for sharing screen and separate window on desktop. Unsure about mobile clients.
Yes, E2EE video call software exists, but then you’d be using something other than Zoom.
Theoretically you could have everyone in a call use a plugin that “added” E2EE to Zoom, but I’m not certain such a plugin exists - and even if it did, ensuring everyone you communicate with Zoom uses it would be enough of a barrier that it’d be simpler to just use something other than Zoom that has E2EE baked in.
Yes. End to end encryption between two parties is pretty trivial to implement. Video and audio is no exception.
I’m not sure about the technicals, but there are some services that are HIPAA compliant, which I assume means something similar to the end-to-end encryption you’re describing. WebeX is the one I know one of my local hospital systems uses.
Wow, wtf…it seems like every big tech company is going through “enshittification”. Is there an open-source alternative for Zoom that is hopefully more privacy-focused?
I haven’t used it but the !selfhosted community has mentioned Jitsy Meet before.
Just discovered it too from !privacy. Reading about it now, it’s FOSS which is always good. Here’s the link: https://meet.jit.si/
There’s element call as well. (https://call.element.io)
Apparently they did:
As of Monday afternoon, [section 10.4] has a new paragraph in bold below it: "Notwithstanding the above, Zoom will not use audio, video or chat Customer Content to train our artificial intelligence models without your consent.”
How exactly they obtain customer “consent” isn’t disclosed.
I’m not so worried about Zoom adding fancy autocomplete (“training our models”) as I am with them licensing it out. This is what section 10.4 says before the caveat:
You agree to grant and hereby grant Zoom a perpetual, worldwide, non-exclusive, royalty-free, sublicensable, and transferable license and all other rights required or necessary to redistribute, publish, import, access, use, store, transmit, review, disclose, preserve, extract, modify, reproduce, share, use, display, copy, distribute, translate, transcribe, create derivative works, and process Customer Content and to perform all acts with respect to the Customer Content, including AI and ML training and testing.
I don’t think that extra caveat even addresses licensing meeting content to third parties for training A.I.
A bit later in the article also addresses this:
That consent, [Zoom Chief Product Officer Smita] Hashim closed, still won’t allow third parties to train an AI off your calls: “And even if you chose to share your data, it will not be used for training of any third-party models.”
However, glancing through the ToS I don’t see where Zoom prohibits third-party AI training, only prohibiting training their own models. On the other hand, data for training LLMs is apparently the modern gold-rush and it’s feasible that Zoom wouldn’t want that data to be accessed by any potential competitors.
you agreed to the TOS, therefore they have your consent.
Holy shit how did I not hear about this but the back to the office thing was everywhere? Fuck that, no. No that is not ok. Fuck. FUCK why is everything going to this shit?
What did zoom change? Other than their workers being forced back to offices
Everything that is exposed to consumer information will be this way soon. It’s an eventuality at this point, not a maybe.
Ha! Joke‘s on them, the Intenet is already runied.
Just as bad, if not worse, than the blockchain craze that came before it.
This is worse. Blockchains didn’t do anything except extract money from suckers, which is bad but has a very limited impact. There were no valid applications, nothing they could do cheaper or better, so they just existed in their own separate world of scammers and grifters. AI is appropriating people’s work to shit up the internet we all use with nonsense and will continue to be used because it has valid applications: creating bad art or vapid text without having to pay anyone.
Blockchains didn’t do anything except extract money from suckers
And consume more energy than Egypt. And drive GPU prices through the roof for a couple of years.
The internet is a place where people can connect from across the world, bringing them closer together and allowing the free exchange of information. It’s unfortunate that it is being more and more locked down and controlled by business and government interests looking to isolate, control, and monetize people. AI isn’t the only tool being used to ruin the internet, but it’s damn effective.
“free exchange” is being capitalized by the rich.
there are countries going through brain drain, this is the internet/world going through a brain drain global recession/dark age
This stuff is only economical at scale, it makes no sense to set up an AI spambot farm on startrek.website because you’ll just get banned and you’ll have to do all the ban evasion stuff you have to do on, say, Facebook but for a site with a hundredth of the daily active users
For now.
Right now, there can be value in using a bot net to create 10000 AI driven bots that look like humans but have a coordinated agenda on Reddit. As that value drops, I expect that smaller and smaller forums will get inundated with these bot networks.
Published on a site with loads of AI-generated articles
It would be the ultimate irony if this article was, itself, generated by AI. Based on the article’s “voice”, I doubt it is, but in this brave new world…one must always wonder.
An AI writing about why it sucks? How human.
Heres another layer; AI arguing about whether an AI written article is misinformation.
deleted by creator
deleted by creator
I feel like with “cheap” AI we’re entering a dark age of tech.
Eventually everything will, the tools and the tech will get more mature and sort itself out but for a couple of months/years we will be confronted with bad AI news , bad AI games, bad AI art, etc
I’m pretty confident a day will come were AI will be seen as a tool and that pure AI generated content will be seen the equivalent of a bad JavaScript game or a cheap knock clone.
