which models did you use and what did you ask? the big and gratis models are very good already. (claude 3.7, openai 4o, deepseek r1). mistrals gratis model is not so good.
A person may not “have something to hide”, but that doesn’t mean they want a digital profile built about them to target for any of a billion reasons. Perhaps someone says something anti-MAGA, and thanks to AI someone gets a list of all that person’s accounts, hobbies, schedules, etc, and that information is then used to target that person.
Everyone in the world has something to hide, especially when AI can aggregate all of those somethings into a human digestible format.
Consider the possibility that you don’t need to be doing anything wrong besides existing to be persecuted by a fascist regime.
SnotFlickerman@lemmy.blahaj.zone
on 28 May 17:23
nextcollapse
Telegram was always a bad actor with misleading claims about their communications service. Not shocking at all.
nightwatch_admin@feddit.nl
on 28 May 17:37
nextcollapse
Oh wow. Integrating an LLM AI in the absolute bottom in human communication; CSAM, drugs, revenge porn, criminal chats, illegal weapons trade and so much more … this will work out well.
We know about all of that bottom human stuff because it’s not encrypted. I guarantee you encrypted messengers like Signal are much worse when it comes to that.
nightwatch_admin@feddit.nl
on 28 May 21:50
collapse
See, believe it or not but as far as we can tell, Signal is just less convenient for large groups and it very much looks like Telegram is worse. But it doesn’t matter: it’s about AI (Grok) being integrated with an environment known for bad data. Grok will interact with bad actors and train on awful data.
sunzu2@thebrainbin.org
on 28 May 18:13
nextcollapse
Hold on.... Why are they paying telegram and not the other way around?!
So you’re telling me that Grok is so bad that they have to pay developers to integrate it? How backwards. People are willing to pay for Grok’s competitors.
threaded - newest
Gross. I’m so sick of these AI tech bros shoving it into every corner of my life.
i really like ai, but only when i can activate it myself in apps or website.
There are valuable uses for AI in general. LLMs are not one of them.
i beg to differ
They’re nothing but manufactured bullshit machines.
which models did you use and what did you ask? the big and gratis models are very good already. (claude 3.7, openai 4o, deepseek r1). mistrals gratis model is not so good.
All of them
It’s probably long past time to remove Telegram from your life.
Most of these are apps are privacy and security vulnerabilities but the normie got nothing to hide
A person may not “have something to hide”, but that doesn’t mean they want a digital profile built about them to target for any of a billion reasons. Perhaps someone says something anti-MAGA, and thanks to AI someone gets a list of all that person’s accounts, hobbies, schedules, etc, and that information is then used to target that person.
Everyone in the world has something to hide, especially when AI can aggregate all of those somethings into a human digestible format.
Consider the possibility that you don’t need to be doing anything wrong besides existing to be persecuted by a fascist regime.
Telegram was always a bad actor with misleading claims about their communications service. Not shocking at all.
Oh wow. Integrating an LLM AI in the absolute bottom in human communication; CSAM, drugs, revenge porn, criminal chats, illegal weapons trade and so much more … this will work out well.
We know about all of that bottom human stuff because it’s not encrypted. I guarantee you encrypted messengers like Signal are much worse when it comes to that.
.
And yet, most radicalized chats are on Telegram, or so we were led to believe.
Isn’t telegram more like Instagram with feeds?
See, believe it or not but as far as we can tell, Signal is just less convenient for large groups and it very much looks like Telegram is worse. But it doesn’t matter: it’s about AI (Grok) being integrated with an environment known for bad data. Grok will interact with bad actors and train on awful data.
Hold on.... Why are they paying telegram and not the other way around?!
To ask is to answer 🐸
…because Grok is being used to push certain “alternative facts” to those who are susceptible.
What’s worse than enshittening?
Telegram
Telegram with Grok.
grok is about to enter its furry terrorist arc
Is it gonna be opt out/opt in
So you’re telling me that Grok is so bad that they have to pay developers to integrate it? How backwards. People are willing to pay for Grok’s competitors.
No, the world’s richest Nazi wants access to all of the communications of Tele-Nazis.