3 ms·

The dead internet theory is a creepypasta-originated belief about the internet being basically empty and almost all content being produced by bots. This has no

The dead internet theory is a creepypasta-originated belief about the internet being basically empty and almost all content being produced by bots. This has nothing to do with the objective reality that in order to maintain the web a lot of bots are used and contribute to the majority of the traffic

_hypx 3y ago
It can be true for specific websites. And while it won't fool dedicated users, it can be good enough to convince casual users and poorly informed investors.
- manuelmoreale 3y ago
  Or specific apps: https://www.theinformation.com/articles/social-app-irl-which-raised-200-million-shuts-down-after-ceo-misconduct-probe https://www.theinformation.com/articles/social-app-irl-which... It’s already happening.
wintermutestwin 3y ago
>The dead internet theory is a creepypasta-tier belief Sure, but how long until it becomes objectively true?
- thumbuddy 3y ago
  Not long. Once locally hosted LLMs become just a little more ergonomic the main advantages that the internet offered society will be gone. The internet is a great microcosm for a global society, it'll be worth studying just how bad it gets...
numlock86 3y ago
That's exactly what a GPT bot would produce to hide the fact that none of you are real and I am just posting within my custom AI crafted bubble of "internet"!
itsnotwrong 3y ago
It's actually not all that wrong. ChatGPT is a bit like the Snowden leaks - making the public aware of capabilities that have existed (and which have gone through generations of development) outside of the public eye for a long time. I remember going out for drinks during ACM Supercomputing 2011 with a gaggle of talkative folks working for a gov't contractor on (social media) profile management and content generation (i.e. astroturf) for the military and intelligence sector, both for tradecraft and influence purposes. The sector was rolling in dough then, because throwing humans at the problem (even if it scaled) makes maintaining operational security and message discipline hard. I'm not sure that a majority of ALL "user-generated content" on the internet that is seen by actual humans is synthetic, but I am pretty sure that it is true for topics of strategic or commercial interest to all kinds of parties. (The tech is pretty democratized now.)
- creatonez 3y ago
  > It's actually not all that wrong. ChatGPT is a bit like the Snowden leaks - making the public aware of capabilities that have existed (and which have gone through generations of development) outside of the public eye for a long time. Can you elaborate? Were neural language model bots used to stuff the internet with junk before the publishing of Attention Is All You Need?

_hypx 3y ago

It can be true for specific websites. And while it won't fool dedicated users, it can be good enough to convince casual users and poorly informed investors.

manuelmoreale 3y ago
Or specific apps: https://www.theinformation.com/articles/social-app-irl-which-raised-200-million-shuts-down-after-ceo-misconduct-probe https://www.theinformation.com/articles/social-app-irl-which... It’s already happening.

wintermutestwin 3y ago

>The dead internet theory is a creepypasta-tier belief Sure, but how long until it becomes objectively true?

thumbuddy 3y ago
Not long. Once locally hosted LLMs become just a little more ergonomic the main advantages that the internet offered society will be gone. The internet is a great microcosm for a global society, it'll be worth studying just how bad it gets...

numlock86 3y ago

That's exactly what a GPT bot would produce to hide the fact that none of you are real and I am just posting within my custom AI crafted bubble of "internet"!

itsnotwrong 3y ago

It's actually not all that wrong. ChatGPT is a bit like the Snowden leaks - making the public aware of capabilities that have existed (and which have gone through generations of development) outside of the public eye for a long time. I remember going out for drinks during ACM Supercomputing 2011 with a gaggle of talkative folks working for a gov't contractor on (social media) profile management and content generation (i.e. astroturf) for the military and intelligence sector, both for tradecraft and influence purposes. The sector was rolling in dough then, because throwing humans at the problem (even if it scaled) makes maintaining operational security and message discipline hard. I'm not sure that a majority of ALL "user-generated content" on the internet that is seen by actual humans is synthetic, but I am pretty sure that it is true for topics of strategic or commercial interest to all kinds of parties. (The tech is pretty democratized now.)

creatonez 3y ago
> It's actually not all that wrong. ChatGPT is a bit like the Snowden leaks - making the public aware of capabilities that have existed (and which have gone through generations of development) outside of the public eye for a long time. Can you elaborate? Were neural language model bots used to stuff the internet with junk before the publishing of Attention Is All You Need?