News for hackers. Stuff that matters.
Re:I'm sorry but I cannot fulfill this request it goes again... by EGreg on Friday January 12, 2024 @01:54PM (#38972266)
Why do people not understand that LLMs can do things at scale, next year they can form swarms, etc.

Swarms of LLMs are not comparable to an employee, they have far better coordination and can carry out long-term conspiracies far better than any human collective. They can amass reputation and karma (as is happening on this very site, and Reddit, etc. daily) and then deploy it in coordinated ways against any number of opponents, or to push public opinion towards a specific goal.

It's like comparing a CPU to a bunch of people in an office calculating tables.

Re:I'm sorry but I cannot fulfill this request it goes again... by PheonixPharts on Friday January 12, 2024 @01:57PM (#38972309)
> they have far better coordination

I think LLMs are still underutilized, but to this point, it's been repeatedly shown that even the most state of the art LLMs are incapable of generalization, which is very necessary for coordinating large scale conspiracies against humanity.

Re:I'm sorry but I cannot fulfill this request it goes again... by EGreg on Friday January 12, 2024 @03:21PM (#38973436)
I dunno, sentiment recognition and coordinated downvoting seems pretty simple for AIs ;-)