Is AI "reading" and commenting on/rehashing some of our threads?

I was doing a web search for my box of 20 thread, linked below, and ran into a page that did "insights" into the thread. It seems to copy some of the wording and then rearrange it or reword things to sound different or offer some added thoughts. It has the feel of some "AI" summaries that I have seen outside of coins.

Does anyone know about this website or if this is someone who is here on CoinTalk?

Original thread:

https://www.cointalk.com/threads/ddddd’s-box-of-20-updated-version-8-0.417618/

New website:

https://coinagerings.org/my-insights-on-ddddds-box-of-20-updated-version-8-0/

 

Log in or Sign up to hide this ad.

If you ever google anything, the top "results?" will be their AI summary. It isn't great, but it has been getting better.

Try searching for a very specific topic that you know has at least a couple CoinTalk threads. You'll be surprised.

And in reference to a recent thread referencing the "bots" on the site... almost certainly, many of the bots are scouring the site for AI.

 

I have seen the google AI summaries and answers to some of the more common coin questions via AI. What I have not seen is someone start their own website to comment about my box of 20 or @lordmarcovan calendar idea.

https://www.cointalk.com/threads/whos-done-a-calendar-with-their-12-caesars-coins.418079/

https://coinagerings.org/my-adventure-crafting-a-calendar-with-twelve-caesars-coins/

 

@physics-fan3.14 they really like you since there are three versions of your post about the PL error :

Original:

https://www.cointalk.com/threads/somebody-made-a-big-oopsie-on-my-newest-prooflike-coin.418242/

The three posts:

https://coinagerings.org/my-big-oopsie-nickel-when-a-prooflike-error-became-a-numismatic-highlight/

https://coinagerings.org/my-big-oopsie-coin-a-prooflike-nickel-with-a-double-strike-surprise/

https://coinagerings.org/somebody-made-a-big-oopsie-on-my-prooflike-nickel-a-numismatic-surprise/

 

@mrweaseluv they enjoyed your post too

https://www.cointalk.com/threads/gtg-1899-morgan.418245/

https://coinagerings.org/my-deep-dive-into-the-gtg-1899-o-morgan-dollar-grading-and-strategy/

 

I don't think I've ever been "plagiarized" before.. even by AI

 

There is a sort of similar thread over on NGC, somebody posting a 1985 "dot" experimental cent, convinced that it's valuable because of the dot, and I'm not sure where experimental came from. He posts an AI response that says the 1985 experimental dot cent is worth thousands, and AI cites "CoinWeek reporting." I go looking for this CoinWeek article about a 1985 experimental dot cent. The only article that I could find is "1985 Lincoln Cent : A Collector’s Guide", and the only reference to a dot and experimental is in a comment on the article! Presumably from the same person posting on NGC ("gene" versus "geno"). So apparently this guy thinks he has something valuable, comments to ask about it on a CoinWeek article that's about 1985 cents in general, then asks AI about it, and AI authoritatively tells him what he wants to hear based on his own post about it.

 

The site was registered in 2017 on GoDaddy. The Internet Archive has a few captures.

Looks like in 2018 it was a business that made rings from silver proof coins in Orlando.

In late 2024 it becomes a parking page. Today it's apparently a botfarm.

 

I've had ChatGPT quote my own comments from CoinTalk in a search for a specific topic. Not helpful for me but it seems CoinTalk is a go-to for AI.

 

Easy to scrape and add to the corpus. Everything that makes it easy to build and operate a board makes it easy to scrape.

 

So is AI what's putting up all the garbage sexually explicit ads too?

 

Noooo... the ads are pretty much based on your browsing history. And, plausible deniability, maybe the browsing history of other users who've visited the same sorts of pages you have.

 

We're being increasingly pressured to make more use of AI at work. A co-worker and I were trying to make some progress on a long-standing issue, and he asked Copilot about it. It promptly came back with a detailed response, citing... a Teams conversation between me and that same co-worker, from several months before, when we had even less idea what was going on.

 

Sounds redundant........

 

Almost all the adds I see on this thread say AI. Disappear and become an AI master. Too funny. AI has its uses but I would certainly double check it.

 

Which is too funny in itself - if you have to double check it, why bother with it. Yeah, I'm a naysayer. If it can't distinguish between good information and bad information, it's useless. Worse than useless, because it's like "it's on the internet so it must be true" on steroids but with a large dose of HAL 9000.