Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

PhysicsForums and the Dead Internet Theory

hallofdreams.org

Post reply on HN
Print view View original

111–120 of 165 posts

Re: PhysicsForums and the Dead Internet Theory

#111

Post by cogman10 »

earnestinger wrote: post #64
Aerroon wrote: post #59

Earlier quoted context omitted.

I can understand it for AI generated text, but I think there are a lot of people that like AI generated images. Image sites like get a ton of people that like AI generated images. Civitai gets a lot of engagement for AI generated images, but so do many other image sites.

I don’t understand the problem with AI generated images. (I very much would like any AI generated text to be marked as such, so I can set my trust accordingly)

> I don’t understand the problem with AI generated images.

Depends on what they are used for and what they are purporting to represent.

For example, I really hate AI images being put into kids books, especially when they are trying to be psuedo-educational. A big problem those images have is from one prompt to the next, it's basically impossible to get consistent designs which means any sort of narrative story will end up with pages of characters that don't look the same.

Then there's the problem that some people are trying to sell and pump this shit like crazy into amazon. Which creates a lot of trash books that squeeze out legitimate lesser known authors and illustrators in favor of this pure garbage.

Quite similar to how you can't really buy general products from amazon because drop shipping has flooded the market with 10 billion items with different brands that are ultimately the same wish garbage.

The images can look interesting sometimes, but often on second glance there's just something "off" about the image. Fingers are currently the best sign that things have gone off the rails.

Re: PhysicsForums and the Dead Internet Theory

#112

Post by sdwr »

StefanBatory wrote: post #47
Aurornis wrote: post #42

Earlier quoted context omitted.

> The forums are still there but not nearly the camaraderie of the early days. I remember visiting those forums when I was young and feeling like part of a big group of friendly people hanging out online together. I tried creating a new account recently and it had a very different vibe. Felt like the old guard had been established and the forums I looked at were dominated by a couple of posters who just wanted to tal…

> Felt like the old guard had been established and the forums I looked at were dominated by a couple of posters who just wanted to talk, but not discuss anything. I think this is the case for most places, I'm afraid. I use mainly Discord - there are certainly a lot of servers where I'm purely because I'm talking to people I met there, and I don't even play that game anymore. There solution is simpler - after time we…

The problem is when the old guard becomes an exclusive clique. Sometimes it's by accident ("I'm happy with the friends I already have"), but usually there's a portion of the inner circle that validate themselves by gatekeeping newcomers.

There has to be an active commitment to include (annoying, tactless, socially-impoverished) newbies, or the snake eats its tail and collapses under its own weight.

Re: PhysicsForums and the Dead Internet Theory

#113

Post by inasio »

firesteelrain wrote: post #87
inasio wrote: post #70

Talk about burying the lede! Near the bottom of the story the site owner confirms that it was him that added the backdated AI comments (perhaps it should have been obvious...)

I couldn’t find it. He was trying to seed the site ?

This is what he replied when asked about the backdated comments:

> The backdated answers were an internal test. We conceived of a bot that would provide a quality answer to a thread without a reply after 1+ years. That too also failed. Instead, I’m considering pruning all threads without a reply as they clutter up the forums.

Re: PhysicsForums and the Dead Internet Theory

#114

Post by hooverd »

scotty79 wrote: post #15
Terr_ wrote: post #3

Ooof. The idea--or reality--that humans' accounts would be hijacked by site-owners to make impersonating slop (presumably to bring in ad-revenue) is somehow both infuriating and energy-sapping-depressing. Issues of trust and attribution have always existed for the web, but for many reasons it feels so much worse now --how bad must it get before some kind of sea-change can occur? I'm not sure what the solution would b…

Shouldn't we invent a protocol that keeps the content you produce under your control so that places like forums or facebook are only discovery devices and interaction facilitators, but not custodians of all communication? Being able to independantly reach the source of piece of information is increasingly important.

That's what ATProto is trying to solve, funny enough.

Re: PhysicsForums and the Dead Internet Theory

#115

Post by bboygravity »

segasaturn wrote: post #17

Money quote: > There’s also a social contract: when we create an account in an online community, we do it with the expectation that people we are going to interact with are primarily people. Oh, there will be shills, and bots, and advertisers, but the agreement between the users and the community provider is that they are going to try to defend us from that, and that in exchange we will provide our engagement and con…

Cool, does that mean we can go out more and talk to real humans again? Can't wait tbh.

Re: PhysicsForums and the Dead Internet Theory

#116

Post by llm_trw »

segasaturn wrote: post #17

Money quote: > There’s also a social contract: when we create an account in an online community, we do it with the expectation that people we are going to interact with are primarily people. Oh, there will be shills, and bots, and advertisers, but the agreement between the users and the community provider is that they are going to try to defend us from that, and that in exchange we will provide our engagement and con…

For the tech discussions I'm interested in burning cpu/GPU cycles for proof of work is a good way to make replies expensive enough that only people who care will post then.

Another option is a web of trust.

It's finally the year of gpg!

Re: PhysicsForums and the Dead Internet Theory

#117

Post by wholinator2 »

layer8 wrote: post #97
lkramer wrote: post #94

Earlier quoted context omitted.

This is interesting. I wonder why it's not visible by default.

Maybe it is/was an A/B test, to see if it hurts ad revenue (it probably does). The option appeared randomly for me on a search, and I took immediately note of the udm number. :)

In addition to the others mentioned on the sibling comment (I cannot reply to it?), udm=28 is shopping, 36 is books, 37 is "products", 44 is "visual matches", 48 is "exact matches", 50 is "AI Mode" but no tab appears, 51 is homework and I stopped at 80 because the page kept removing that part of the url all the way up to that point.

Re: PhysicsForums and the Dead Internet Theory

#119

Post by nxobject »

COAGULOPATH wrote: post #27

Something I'm increasingly noticing about LLM-generated content is that...nobody wants it. (I mean "nobody" in the sense of "nobody likes Nickelback". ie, not literally nobody.) If I want to talk to an AI, I can talk to an AI. If I'm reading a blog or a discussion forum, it's because I want to see writing by humans . I don't want to read a wall of copy+pasted LLM slop posted under a human's name. I now spend dismayin…

Another trick I do is to scroll to the end, and see if the last paragraph is written as a neat conclusion with a hedge (i.e. "In short...", "Ultimately..."). I imagine it's a convention to push LLMs to terminate text generation, but boy is it information-free.

Re: PhysicsForums and the Dead Internet Theory

#120

Post by nxobject »

agumonkey wrote: post #67
COAGULOPATH wrote: post #27

Something I'm increasingly noticing about LLM-generated content is that...nobody wants it. (I mean "nobody" in the sense of "nobody likes Nickelback". ie, not literally nobody.) If I want to talk to an AI, I can talk to an AI. If I'm reading a blog or a discussion forum, it's because I want to see writing by humans . I don't want to read a wall of copy+pasted LLM slop posted under a human's name. I now spend dismayin…

nobody wants to see other's ai generated images, but most people around me are drooling about generating stuff wait for the proof-of-humanity decade where you're paid to be here and slow and flawed

Re: proof-of-humanity... I'm looking forward to a Gattaca-like drop-of-blood port on the side of your computer, where you prick yourself everytime you want a single "certified human response" endorsement for an online comment.
Post reply on HN