AI writing messages on Bitcointalk.org

[Lucasgabd]

I agree @Welsh
I'm basing my assumptions on the ideas that translation got much better 10 years after mvp, and chatGPT is the first mvp released to the public

the dataset possibility for AI ends up being the whole internet
which will keep growing and have more data with time

low level jobs like junior copywriters and even designers are probably doomed with chatGPT and Dall-e if they don't adapt

on the other hand the amount of trash content we'll see generated in blogs and social media will be higher than we've ever seen

[1714959551]

1714959551

[DdmrDdmr]

<…> the dataset possibility for AI ends up being the whole internet <…>

Not quite. Apparently, the model was trained on data obtained by resorting to some scraping services, which provide downloadable data formats of the scraped data (I haven’t looked into the details too much), alongside some book collections and Wikipedia:

<…>
I’ve tried to find out how the training data for the model behind Chat GPT, which is not easy to find in detail. The closest I’ve managed to come-across is the description provided here.


It’s better to follow the links provided by the content author is one really wants to get a broarder idea of what’s behind the summarized table (i.e. Book1 is Bookcorpus, which includes over 11k books; Common Crawl is a non-profit organization that provided petabytes of scraped data that goes as far back as 2011, and so forth).

Note:
Interesting/non-interesting surprise: Checkout the names of the founders of Open AI back in 2015.
Clue: Twitter’s current (for now) CEO

I hadn't heard of Common Crawl crawl before, figuring as a massive source of data for these types of projects when you look at it.

[LoyceV]

There's also something called Moore's law. Meaning, we have exponential growth because we don't know a lot about a subject at first, but then that starts to plateau the more we know about it. Translations, and AI are very likely subject to this just as much as computers, and performances. So, while we might expect it to get better, determining how much it might get better is the difficult part.

Eventually, once enough AIs spam the internet, they'll mostly get data from other AIs instead of humans as input. By then the internet will be one big circle jerk of AIs trying to impress each other, and us mortals move on to something else.

[Pmalek]

Humans make mistakes. Grammar, typos, thinking errors like brain farts. You cannot really replicate those tiny human imperfections perfectly and they are what makes us what we are (yet).

Humans make grammatical, semantical, and punctuation errors, but a simple tool like Grammarly can detect all that and suggest what needs to be corrected. And there is nothing wrong with wanting your posts to sound and look good. Besides Grammarly, all popular browsers have spellchecks that help if you misspell something or make a typo. There is also nothing wrong in correcting such mistakes, so you don't sound like a 9-year old who can't spell. Couple the two tools together, and you get an almost perfect-like poster.   

[AverageGlabella]

Humans make mistakes. Grammar, typos, thinking errors like brain farts. You cannot really replicate those tiny human imperfections perfectly and they are what makes us what we are (yet).

Humans make grammatical, semantical, and punctuation errors, but a simple tool like Grammarly can detect all that and suggest what needs to be corrected. And there is nothing wrong with wanting your posts to sound and look good. Besides Grammarly, all popular browsers have spellchecks that help if you misspell something or make a typo. There is also nothing wrong in correcting such mistakes, so you don't sound like a 9-year old who can't spell. Couple the two tools together, and you get an almost perfect-like poster.   

I do not know if grammarly supports the german language but most translators get the grammar wrong in German a lot of the time it is easy to spot someone that is using a translation tool because the grammar is just not right. I think this is true for most spell checkers they are only good for popular languages like English where there is a lot of people correcting the mistakes. Most grammar and spell checkers use American English too so it might not be perfect if you are in a country that does not use american english as the spelling would be different for certain words.

[Pmalek]

I do not know if grammarly supports the german language...

Grammarly only works with English and I think its default settings is British English, not American.   

Most grammar and spell checkers use American English too so it might not be perfect if you are in a country that does not use american english as the spelling would be different for certain words.

Unless you are a language expert or a native that pays attention to that sort of thing, it's not something you will notice. Surely those that rely heavily on spellcheckers wouldn't mind which variant they work with. Your browser settings can be configured to use spellchecking for British or American English, so that's adjustable according to your needs and isn't a problem to change.

[Synchronice]

Remember, it uses the information out there. Basically, we're feeding it a massive data set already, and without improving that data set, it's unlikely to improve massively, as the dataset is the most important factor for generating this sort of content.

Yeah, it uses information out there but if it collects all the data that has been published on the web, including posts on forums, articles from news websites, published pdf books, etc and then makes a serious analyze to develop itself, then I think that it will be able to even evolve from its own massive data. It can use the past to predict the future and create tons of variations. Well, I don't want to go deep further, just want to point out that everything generated AI is meaningless and I don't even know why we should focus on it. Why the fuck would someone read AI generated articles? What's the purpose? What's the point? I don't understand what kind of joy can someone find on artificially generated machine that follows manually written logic and rules.

[virasog]

Yeah, it uses information out there but if it collects all the data that has been published on the web, including posts on forums, articles from news websites, published pdf books, etc and then makes a serious analyze to develop itself, then I think that it will be able to even evolve from its own massive data.

Since this ChatGPT have all the data stored in its database from different sources already present on the internet, would it be possible that it may produce plagiarized content?
Have anyone checked the content produced by Chat GPT on plagiarism checker tools to confirm if it always produces a unique article?

It can use the past to predict the future and create tons of variations.

I don't think this chat GPT cannot predict the future because it has only data till 2021. For example, today I asked him who won the football world cup 2022, on which he has no clue because his database is only till 2021.

[LoyceV]

Have anyone checked the content produced by Chat GPT on plagiarism checker tools to confirm if it always produces a unique article?

I have no doubt the article as a whole is unique, but that doesn't mean it's not plagiarism. Look at the definition:

Plagiarism is presenting someone else’s work or ideas as your own, with or without their consent, by incorporating it into your work without full acknowledgement. All published and unpublished material, whether in manuscript, printed or electronic form, is covered under this definition. Plagiarism may be intentional or reckless, or unintentional. Under the regulations for examinations, intentional or reckless plagiarism is a disciplinary offence.

This is literally what this AI computer system does.

[Aliev]

Have anyone checked the content produced by Chat GPT on plagiarism checker tools to confirm if it always produces a unique article?

Traditional plagiarism checkers like Grammarly, Quetext, or Copyscape can't tell if it's AI-generated or not. They usually think it's 100% plagiarism-free.

However, people are already developing tools that detect specifically AI content, and they're showing pretty accurate results. Now it's not about hunting down plagiarism, it's now about detecting AI-generated content.

[dkbit98]

I don't think this chat GPT cannot predict the future because it has only data till 2021. For example, today I asked him who won the football world cup 2022, on which he has no clue because his database is only till 2021.

This just shows how stupid this chatGPT really is, if there are no humans to feed him new information and training all the time, he would soon become irrelevant junk.
They are probably going to start asking money to feed this computer new data in near future, and people playing with this crap are doing it for free now.
Repeating stuff and mixing words doesn't make something intelligent, and there is always some human ^{(good or evil)} who is controlling this machine.

[Lucius]

I have one doubt about whether the post written by AI is something that should be tolerated on the forum, even if the OP admitted that the post was created that way? I am specifically interested in this case -> https://bitcointalk.org/index.php?topic=5432035.0 (archived)

[FatFork]

I have one doubt about whether the post written by AI is something that should be tolerated on the forum, even if the OP admitted that the post was created that way? I am specifically interested in this case -> https://bitcointalk.org/index.php?topic=5432035.0 (archived)

I think it's important to remember that the purpose of a forum is to foster meaningful and respectful discussion among its members. Therefore, there is a line that should be drawn when it comes to content that is not genuine or authentic. Ultimately, it's up to the moderators and policies of the forum to decide if they want to put up with artificial content, but in my opinion, it doesn't belong on a bitcointalk forum. It's like trying to have a conversation with a robot – it's just not the same as talking to a real person. 

[Don Pedro Dinero]

I have one doubt about whether the post written by AI is something that should be tolerated on the forum, even if the OP admitted that the post was created that way? I am specifically interested in this case -> https://bitcointalk.org/index.php?topic=5432035.0 (archived)

I think it's important to remember that the purpose of a forum is to foster meaningful and respectful discussion among its members. Therefore, there is a line that should be drawn when it comes to content that is not genuine or authentic. Ultimately, it's up to the moderators and policies of the forum to decide if they want to put up with artificial content, but in my opinion, it doesn't belong on a bitcointalk forum. It's like trying to have a conversation with a robot – it's just not the same as talking to a real person.  

Of course it should be allowed, especially since the OP did not hide and made it explicit that the text was written by an IA. I can't imagine on what grounds such posts could be censored here when they are allowed all over the internet.

And I came to comment on that post myself because it shows that the IA has a lot of room for improvement, because it wrote a post about how to make money in gambling that is rubbish. It is clear that it has taken information from here and there, but it ignores the main thing, that you can not make money in gambling unless we are talking about specific games such as sports betting or poker, which the text does not do.

[FatFork]

Of course it should be allowed, especially since the OP did not hide and made it explicit that the text was written by an IA. I can't imagine on what grounds such posts could be censored here when they are allowed all over the internet.

Specific reasons for censoring AI generated posts may vary depending on the specific forum and its rules and policies. For example, someone can argue that such posts are spammy, zero or low value, misleading, etc... Btw, it's important to note that AI-generated content is not universally accepted on the Internet. In fact, sites like Reddit have specific rules in place to prohibit the use of such content.

And I came to comment on that post myself because it shows that the IA has a lot of room for improvement, because it wrote a post about how to make money in gambling that is rubbish. It is clear that it has taken information from here and there, but it ignores the main thing, that you can not make money in gambling unless we are talking about specific games such as sports betting or poker, which the text does not do.

And, how do you see this kind of discussion related to Gambling?
I want to point out that the topic of AI-generated content was created in the gambling section, but you are discussing the quality and accuracy of AI-generated content. That discussion is not related to gambling and might be more suitable for a different forum board.

[Lucius]

I was just wondering what other people's attitude is when it comes to such posts, because it is not acceptable to me personally - the same as any post that is simply copy-pasted and a link is added as a source. Tolerating this kind of practice would actually mean that anyone can use AI for their posts, and I agree with @FatFork that we don't want discussions with robots.

It seems that the shitposters in the Gambling board don't mind it at all, and the OP was even awarded with a merit - the post has been reported, we are waiting for the decision of the mods.

[Accardo]

A.I contents is a restructure of the original article or content. If used in the forum can easily be detected by spam trackers. Unless the user only posts threads, but if used for responses, I don't think A.I uses synonyms, the A.I writer is required to copy someone's comment and shuffle the words to change the context. Before responding. When concretly read the reader can detect easily that they've read such information and trace back to the original response. A.I is mainly used by bloggers and it's not worth to be used on forums.

[Rikafip]

Of course it should be allowed, especially since the OP did not hide and made it explicit that the text was written by an IA. I can't imagine on what grounds such posts could be censored here when they are allowed all over the internet.

Just because something is allowed in other places doesn't mean that it should be allowed too here as well. Take for example stance that this forum his on plagiarism, compared to all the other forums. And yeah, that post should be dealt in the same way mindless copy/paste is (mp matter the source link), meaning deletion.

It seems that the shitposters in the Gambling board don't mind it at all, and the OP was even awarded with a merit - the post has been reported, we are waiting for the decision of the mods.

An average post quality in that board is so low that the post in question is above average one and every new topic is seen as an opportunity to fill signature quota. Not that situation is much better in other parts of the forum, but there is the most obvious.

[LoyceV]

I was just wondering what other people's attitude is when it comes to such posts, because it is not acceptable to me personally

I'm starting to suspect some posts are AI-written, and I don't like it. It's just the next phase of shitposting.
Example: this post: it has a lot of "generic text", and some parts are just bad practice and shouldn't be advised. All in all, it's pointless. But it's well-written and not against the forum rules.

[Little Mouse]

I'm starting to suspect some posts are AI-written,

Same here. I have also started to suspect the same. I have been monitoring one account for the last few weeks, and within this time, the account have posted 3/4 times only which seems like AI generated to me. I don't want to publicly post that until I'm sure that he is posting AI generated text.
This will be somewhat a disaster if someone with a long list of account starts to participate in signature campaign and post AI generated garbage, I'm pretty much sure we will see the day in the coming future.