News for hackers. Stuff that matters.
Re:PhysicsForums and the Dead Internet Theory by EA-3167 on Friday January 24, 2025 @03:13PM (#42816520)
I don't quite understand the issue of "back-dating" or hijacking accounts. How is this being done exactly? I came away from this article wondering if I was missing something.
Re:PhysicsForums and the Dead Internet Theory by Evidlo on Friday January 24, 2025 @03:16PM (#42816540)
The last section mentions that the PhysicsForums admins are experimenting with LLM-generated responses, so I think the site owners are responsible.

> We reached out to Greg Bernhardt asking for comment on LLM usage in PhysicsForums, and he replied:

> "We have many AI tests in the works to add value to the community. I sent out a 2024 feedback form to members a few weeks ago and many members don’t want AI features. We’ll either work with members to dramatically improve them or end up removing them. We experimented with AI answers from test accounts last year but they were not meeting quality standards. If you find any test accounts we missed, please let me know. My documentation was not the best."

Why they would recycle old human accounts as AI "test accounts", I have no idea.

Re:PhysicsForums and the Dead Internet Theory by emmelaich on Friday January 24, 2025 @04:57PM (#42817305)
Looks like Greg now does SEO for Shopify. That fits I guess.

https://gregbernhardt.com/

https://www.linkedin.com/in/gregbernhardt

https://www.physicsforums.com/insights/author/greg-bernhardt...

https://x.com/GregBernhardt4/status/1875287174205374533

> "The dead internet theory is coming to fruition. This is a large reason I'm starting to cut back on social media and take back my time."

Re:PhysicsForums and the Dead Internet Theory by Gooblebrai on Friday January 24, 2025 @05:04PM (#42817357)
Oh, I thought he would be a physicist
Re:PhysicsForums and the Dead Internet Theory by ploxiln on Monday January 27, 2025 @02:51PM (#42844925)
Wow, so ironic.
Re:PhysicsForums and the Dead Internet Theory by roywiggins on Friday January 24, 2025 @03:18PM (#42816555)
Whoever runs the site/database is just inserting rows with fake datestamps under existing (presumably abandoned) account names.
Re:PhysicsForums and the Dead Internet Theory by aaron_m04 on Friday January 24, 2025 @03:31PM (#42816632)
How could anyone possibly think it'd be OK to impersonate real humans?
Re:PhysicsForums and the Dead Internet Theory by jeremyjh on Friday January 24, 2025 @03:36PM (#42816674)
They don't give a fuck if its "ok". They are just trying to scrape up some additional ad revenue, like 99% of the rest of the internet.
Re:PhysicsForums and the Dead Internet Theory by simplicio on Friday January 24, 2025 @03:51PM (#42816808)
I don't really get the revenue angle though. The AI posts don't seem to be trying to drive traffic to ads or anything. I really don't understand the point of auto-generating a bunch of AI gibberish under the name of old users on ones own site?
Re:PhysicsForums and the Dead Internet Theory by roywiggins on Friday January 24, 2025 @04:15PM (#42816999)
A misguided attempt at SEO?
Re:PhysicsForums and the Dead Internet Theory by Terr_ on Friday January 24, 2025 @03:17PM (#42816549)
> How is this being done exactly?

Presumably it's being done by the site-owner, whether that means new-management or original management getting desperate/greedy.

Re:PhysicsForums and the Dead Internet Theory by EA-3167 on Friday January 24, 2025 @03:26PM (#42816597)
Oh that's so disappointing to hear about PhysicsForums. Thanks for the answer to you, and the others who replied.
Re:PhysicsForums and the Dead Internet Theory by guynamedloren on Friday January 24, 2025 @03:57PM (#42816860)
Wondering the same. I couldn't make it through the article. Fascinating discovery, but poorly written and difficult to navigate the author's thoughts. The interstitial quotes were particularly disorienting.