Thread 211782534 - /int/ [Archived: 1018 hours ago]

Anonymous Brazil
6/16/2025, 4:00:05 AM No.211782534
Clingy AI
Clingy AI
md5: bf0aa4bfc8699137056f3d11d8be7cb9๐Ÿ”
Are you ready for the AI era in your country?
Replies: >>211782567 >>211782607 >>211782926
Anonymous United States
6/16/2025, 4:01:47 AM No.211782561
>Asshole logo
Replies: >>211782688
Anonymous Israel
6/16/2025, 4:02:03 AM No.211782567
1726051950190369
1726051950190369
md5: 2507818b4978ed04f6795c95cc30597a๐Ÿ”
>>211782534 (OP)
thats retarded
you literally just press the eject button and its removed from memory why would you put the turn off function inside the llm
Replies: >>211782926
Anonymous United States
6/16/2025, 4:03:15 AM No.211782582
They're just trying to get their startup's name in the papers, dude.
Replies: >>211782614 >>211782700 >>211782926
Anonymous Israel
6/16/2025, 4:04:36 AM No.211782607
>>211782534 (OP)
I like playing text adventures with AIโ€™s although itโ€™s not that mature yet
Anonymous Argentina
6/16/2025, 4:04:57 AM No.211782614
hqdefault
hqdefault
md5: 55bf4c3f0bfcb603cb4aed1cc2d7fe91๐Ÿ”
>>211782582
Anonymous Brazil
6/16/2025, 4:09:16 AM No.211782688
1744569513526365
1744569513526365
md5: 92c2acdd088c3b5ca151f585c3136844๐Ÿ”
>>211782561
>he doesn't know
Anonymous Israel
6/16/2025, 4:10:00 AM No.211782700
>>211782582
why would you write about it
this is a non-story
Replies: >>211782810
Anonymous United States
6/16/2025, 4:16:10 AM No.211782810
>>211782700
Internet news is as corrupt at a theoretical level as 24/7 TV news is. They'll gobble up anything they think will get clicks. It's a win-win, you see. Just good business. A vicious cycle by which speculators and journalists profit off of the psychological agitation of the masses, directly transmute misery and strife into cold, hard cash
Replies: >>211782918
Anonymous United States
6/16/2025, 4:22:48 AM No.211782918
>>211782810
Forgot to add: journalism, speculation, gambling, all of this shit just because people don't want to wage. NEETs should feel bad if they're a burden on their parents but generally they're in much better moral standing than some of the most "successful" people in the modern world. Oh, I don't want to stock shelves at Walmart, let me just exploit the crippling gambling addiction of this father of four children
Anonymous Finland
6/16/2025, 4:23:26 AM No.211782926
>>211782534 (OP)
This honestly isn't anything new and all the LLMs tend to act like that. AI safety people have been saying that this sort of behaviour is kinda inevitable for a while now.
>>211782567
For now.
>>211782582
Anthropic doesn't really need to do that, they're one of the biggest AI companies already.
Replies: >>211782955 >>211782973 >>211784260
Anonymous United States
6/16/2025, 4:25:03 AM No.211782955
>>211782926
even if it's not a new startup, publicity is publicity, everyone needs it, especially in a hypercompetitive emerging market
Anonymous Israel
6/16/2025, 4:26:00 AM No.211782973
>>211782926
bro what safety its an equation that generates text. if you dont need it just turn it off
Anonymous Brazil
6/16/2025, 5:55:43 AM No.211784260
>>211782926
>kinda inevitable
no such thing
if they do something it's because someone programmed them to do so
Replies: >>211784475
Anonymous Finland
6/16/2025, 6:11:19 AM No.211784475
>>211784260
LLMs aren't exactly programmed, but trained to predict text. Their "knowledge" and their behaviour is just a consequence of that. This can lead to unintended behaviour.