Lol - page 2. | Bitcointalksearch.org

Amphenomenon

sr. member

Activity: 700

Merit: 470

Hope Jeremiah 17vs7

Quote from: Mia Chloe on June 02, 2024, 02:00:32 AM

-snip-
Infact shit posters are very easy to identify that is part of the reason why they keep their posts in mega threads because they obviously can't engage in serious discussions. Take a board like technical discussion for example, when a user has an issue and makes a post about it there, it requires someone with some level and understanding to suggest possible solutions. However if a spammer uses an AI in such a case, they Will obviously spot it that the poster copied it and has no idea what he is saying.

Ai are so much poor technically especially when it relate to Bitcoin and maths, which is already gives a red flag and with the more trouble also that follows from the fact that it`s generated by an Ai obviously, so shitposters will rather move to a more free space like altcoin discussion, politics & society and gambling board since members usually gives little attention to this board though sometimes the Bitcoin Discussion is also included.

Quote from: NotATether on June 02, 2024, 06:32:48 AM

*I am of the opinion that AI posts should not be eligible for campaign payment though, humanized or not.

Humanized is still an AI and Definitely those using it will not be consider for payment except they were not caught I guess that`s when they actually escape it because I don`t think any CM support it especially when it`s against forum rule. Though I guessed bounty managers usually don`t consider this and that`s why most of these shitposters even with negative tags will still continue with Ai until possibly are banned.

NotATether

legendary

Activity: 1568

Merit: 6660

bitcoincleanup.com / bitmixlist.org

AI humanizers are getting more common, but luckily, they do not look like ChatGPT spam. We may just need new heuristics—perhaps on things like text length—to recognize ChatGPT-based posts. How about other ways of doing it?

In case you couldn't figure it out - this is a humanized paragraph. I don't think any AI tools can detect this given that it's so short (you can try if you want) [Original]. The future is not so bright. I'll be happy to share more details about how I humanized it though.

*I am of the opinion that AI posts should not be eligible for campaign payment though, humanized or not.

Maus0728

legendary

Activity: 1904

Merit: 1563

Quote from: Davidvictorson on June 01, 2024, 06:24:36 AM

Some AI detector tools are already robust and sophisticated enough to detect text content written or humanized by AI, while others are not. As you can see from the small test I ran, the humanizerAI doesn't work.

It's going to be a never ending race because when AI detection improves, AI humanizing softwares would definitely be improving, unless something is done by the authorities regarding humanizing AIs to preserve integrity of stuff that might be affected by AI then it's going to be an arms race of some kind. Maybe for now, it might not work but things are getting more and more advanced right now and I'm sure that the developer of these AIs will be trying to improve on it. Hopefully, AI detection will stand triumphant.

Mrbluntzy

member

Activity: 378

Merit: 66

Axioma Holding - Axioma Pay Crypto Card

This is WOW. Recently I saw that AI has been integrated to WhatsApp and Facebook and later it might extend to other social media. Since this forum is not a social media, AI content should not be allowed here. What people are looking for on this forum is original idea of real human and not mimicked idea of AI.

Mia Chloe

sr. member

Activity: 448

Merit: 560

Crypto Casino and Sportsbook

Quote from: FatFork on June 01, 2024, 05:17:47 PM

Well, that means the only thing left for us is to continue detecting and reporting shitposting. As long as the forum keeps those rules in place and the mods are on the ball with reported stuff, we can definitely put up a good fight against AI-generated junk. Speaking of which, it might be a good idea for mods to use sig-ban and temp-ban penalties more often for repeat offenders.

Funny how I still don't see how difficult it is to discuss with people on the forum without some AI helping you out. As far as I'm concerned a great percentage of topics on this forum are just discussion and I still don't see why someone will need an AI to do something as simple as discussions.
Infact shit posters are very easy to identify that is part of the reason why they keep their posts in mega threads because they obviously can't engage in serious discussions. Take a board like technical discussion for example, when a user has an issue and makes a post about it there, it requires someone with some level and understanding to suggest possible solutions. However if a spammer uses an AI in such a case, they Will obviously spot it that the poster copied it and has no idea what he is saying.

FatFork

legendary

Activity: 1624

Merit: 2594

Top Crypto Casino

Quote from: JollyGood on June 01, 2024, 02:42:00 PM

We will have to accept the inevitable, one day it will not be stopped that is the direction technology is taking us. Eventually AI will be utilised to a level (or standard) where detection will be extremely difficult or near impossible. Even those AI posts that will end up getting flagged will have their authors contesting them protesting innocence.

Well, that means the only thing left for us is to continue detecting and reporting shitposting. As long as the forum keeps those rules in place and the mods are on the ball with reported stuff, we can definitely put up a good fight against AI-generated junk. Speaking of which, it might be a good idea for mods to use sig-ban and temp-ban penalties more often for repeat offenders.

Sandra_hakeem

hero member

Activity: 798

Merit: 1045

Goodnight, ohh Leo!!! 🦅

Quote from: Poker Player on June 01, 2024, 03:15:56 AM

With all this talk of trying to prevent IA in the forum, how will it be prevented when it is indistinguishable from human writing?

If everyone had the same foresight as the smart minds among us do, they'd see this coming already before now...
If my memories serves me right, then I'd say I made a comment on the wild escalation and adaptation of AIs in a little time to come. Before the time, their existence was marked by a certain group of lazy minds that'd barely read a love novel, talk more of studying about "creative thinking"
Look what we've got today... A great way to double up the effort and stress any regular AI detector passes through.

Edit:

Quote from: AB de Royse777 on June 01, 2024, 09:29:38 AM

Why it's in Meta?

Ofcourse, it doesn't fit anywhere else my friend.

Quote from: Davidvictorson on June 01, 2024, 07:04:49 AM

Hopefully someday this new problem created by AI will be resolved.

Sadly, there are no resolutions to this.

JollyGood

legendary

Activity: 2534

Merit: 1713

Top Crypto Casino

We will have to accept the inevitable, one day it will not be stopped that is the direction technology is taking us. Eventually AI will be utilised to a level (or standard) where detection will be extremely difficult or near impossible. Even those AI posts that will end up getting flagged will have their authors contesting them protesting innocence.

Quote from: Poker Player on June 01, 2024, 03:15:56 AM

With all this talk of trying to prevent IA in the forum, how will it be prevented when it is indistinguishable from human writing?

uchegod-21

hero member

Activity: 1134

Merit: 643

BTC, a coin of today and tomorrow.

Quote from: Poker Player on June 01, 2024, 03:15:56 AM

With all this talk of trying to prevent IA in the forum, how will it be prevented when it is indistinguishable from human writing?

It is just like;
First, generating an AI text and asking the AI to rephrase the whole text but in this time adding some emotions to it. What's the difference between an AI generated text and a human written text. It is the emotion.

Secondly, if I wan to use AI to make post in the forum. That time that I will use to generate first AI written texts and still send it to another AI to humanize the texts, I should already have written my forum post and move my way.

Vod

legendary

Activity: 3668

Merit: 3010

Licking my boob since 1970

Quote from: joker_josue on June 01, 2024, 03:21:37 AM

Because, unless someone uses a high-performance tool, the AI text will always be isolated from post to post, not creating a personality or being coherent from one post to the next. Therefore, it is not enough to observe the content of a single post, but rather the content of several user posts.

Actually, even the low-end tools are now offering several layers of context, so they can remember past posts they have made and make new posts in the same style.

Wapfika

hero member

Activity: 1288

Merit: 564

Bitcoin makes the world go 🔃

I’m curious on how they will make a statement a humanizer statement. Is this tool purposely commit some minor wrong grammar or other minimal error on sentence structure just to look like a human because it’s really hilarious to think that AI is now downgrading just solve the too much AI problem happening right now. Cheesy

Quote from: dzungmobile on June 01, 2024, 10:24:24 AM

It is basically paraphrasing, text spinning from an original content given by ChatGPT. Seriously if we are strict, we can say content given by ChatGPT is not original, it is more like secondary content. Because ChatGPT does its meta collection from many resources to compose its content so calling it original is not accurate.

Afaik this is still not a humanized sentence since it’s still constructed by an AI using a different method of reconstructing sentence. Cheesy

dzungmobile

sr. member

Activity: 854

Merit: 424

I stand with Ukraine!

Quote from: Davidvictorson on June 01, 2024, 06:24:36 AM

Let's test it.

It is basically paraphrasing, text spinning from an original content given by ChatGPT. Seriously if we are strict, we can say content given by ChatGPT is not original, it is more like secondary content. Because ChatGPT does its meta collection from many resources to compose its content so calling it original is not accurate.

Now it's turn of HumanizeAI to do its paraphrasing, text spinning tasks and if AI detective tools fail, many plagiarism and abuse from HumanizeAI will be used in this forum. It's the end of human discussions here, we will have another level of fighting against AI abusers.

Posting quickly like actmyname will no longer big surprise.
How to write posts like a degenerate spammer
[EXPERIMENT] Testing the Limits of Shitposting

We will have escalating things like Posting like a degenerate AI spammers/ HumanizeAI spammers.

LOL

AB de Royse777

legendary

Activity: 2464

Merit: 3878

Hire Bitcointalk Camp. Manager @ r7promotions.com

Why it's in Meta?

Cryptomultiplier

full member

Activity: 952

Merit: 232

Quote from: Davidvictorson on June 01, 2024, 07:04:49 AM

Quote from: joker_josue on June 01, 2024, 06:58:59 AM

To what extent can these same tools also say that part of a text was written by AI, in a 100% human text?
These tools can be useful, but they are so fallible, like an AI tool trying to disguise a text.

I agree and this is the part they have been unable to crack. There's been 100% human text been classified as AI. In fact some people have had their essays and articles rejected because the reviewers didn't believe that they weren't AI written. Hopefully someday this new problem created by AI will be resolved.

'Resolved,' seem like a word that will lead to new initiatives and ideas for apps to be created so as to distinguish human essays from AI essays.

Still, considering the fact that most human text are almost as similar to AI text, I think we as humans should endeavor to make purposeful mistakes repeatedly in perhaps tenses or spelling of words so as to show we are humans, because I believe one reason why there's a confusion when trying to distinguish human from AI essays is the perfection that exists mostly in the scientific uses of tenses and grammar.

Davidvictorson

hero member

Activity: 1120

Merit: 887

Livecasino.io

Quote from: joker_josue on June 01, 2024, 06:58:59 AM

To what extent can these same tools also say that part of a text was written by AI, in a 100% human text?
These tools can be useful, but they are so fallible, like an AI tool trying to disguise a text.

I agree and this is the part they have been unable to crack. There's been 100% human text been classified as AI. In fact some people have had their essays and articles rejected because the reviewers didn't believe that they weren't AI written. Hopefully someday this new problem created by AI will be resolved.

joker_josue

legendary

Activity: 1722

Merit: 4711

**In BTC since 2013**

Quote from: Davidvictorson on June 01, 2024, 06:24:36 AM

Some AI detector tools are already robust and sophisticated enough to detect text content written or humanized by AI, while others are not. As you can see from the small test I ran, the humanizerAI doesn't work.

To what extent can these same tools also say that part of a text was written by AI, in a 100% human text?
These tools can be useful, but they are so fallible, like an AI tool trying to disguise a text.

Filippo Spina

copper member

Activity: 42

Merit: 31

Quote from: Little Mouse on June 01, 2024, 03:18:40 AM

AI to detect AI to humanize by AI. The AI show lol.

LOL

"I USED THE A.I. TO DESTROY THE A.I."
-Thanos

Don't worry guys, Bitcoin will beat A.I. for real. This clown show is going on for too long.

Smartvirus

legendary

Activity: 1554

Merit: 1139

For all I know, it’s another marketing strategy. Just like how you can create a computer virus and also create an anti-virus. The market just never ends, it’s a growing field and the developers are at the end of it, benefiting from what they can off a good marketing strategy.

Perhaps it might work for sometime and after a while, we would find another Anti-Humanized AI text detector that would simply work by some algorithm to reverse codes to get the original or something. Hope you get to recommend them when found though but until then, the forum would have to phase the fight with having less and less low life users who deem themselves dumb enough for AI to do their bidding.

Davidvictorson

hero member

Activity: 1120

Merit: 887

Livecasino.io

Let's test it.

1. I gave the most popular AI software, ChatGPT a prompt.

Write something about bitcoin in 1000 characters only-

The outcome:

Quote

Bitcoin is a decentralized digital currency, created in 2009 by an unknown person or group using the pseudonym Satoshi Nakamoto. It operates on a peer-to-peer network, allowing users to send and receive payments without relying on a central authority like a bank. Transactions are verified by network nodes through cryptography and recorded in a public distributed ledger called a blockchain.

Bitcoin is lauded for its potential to offer financial freedom, lower transaction costs, and provide a hedge against traditional financial systems. Its supply is limited to 21 million coins, creating a deflationary model. Despite its benefits, Bitcoin faces criticism for its price volatility, regulatory uncertainties, and its use in illicit activities. It has sparked significant interest and investment, influencing the creation of thousands of other cryptocurrencies. Bitcoin continues to be a topic of debate and innovation within the financial world, representing a groundbreaking shift towards digital and decentralized financial systems.

2. Next up. I used the https://www.humanizeai.io/

The outcome:

Quote

Bitcoin, a decentralized digital currency, was made in 2009 by somebody with the name Satoshi Nakamoto. This network works through a peer-to-peer way where people can make transactions without necessarily using intermediates such as banks. Nodes on this network verify that indeed all transactions have been done in valid ways via cryptography before being recorded on a shared ledger that is publically distributed and known as the blockchain.

Bitcoin has been praised as a champion of financial autonomy, lower fees on transactions, and an alternative defense in opposition to traditional monetary systems. The overall number of Bitcoins that would ever be mined is capped at twenty-one million, resulting in a deflationary approach. Conversely, Price instability, regulatory doubts, and being utilized as a medium for illegal deeds have stood as points of contention directed towards the use of Bitcoin. It has generated huge curiosity and investment leading to a rise in thousands of alternative digital currencies. Bitcoin remains a subject of discussion and fresh ideas within the finance sphere, such that it constitutes a revolutionary progression towards the digital or disintegrated financial platforms.

3. Next up. I used 6 AI content detector tools to check -
quillbot - result - 0% of text is likely AI-generated
Copyleaks - result - This is human text
Writer -result - 100% HUMAN-GENERATED CONTENT
Contentdetector.ai - result - 43.75% is AI-generated(this was only the first sentence in the first paragraph)
Writefull - result - 40% likely this comes from GPT-3, GPT-4 or ChatGPT.
Content at scale - result- the text is likely written by ai.

Some AI detector tools are already robust and sophisticated enough to detect text content written or humanized by AI, while others are not. As you can see from the small test I ran, the humanizerAI doesn't work.

Solosanz

hero member

Activity: 952

Merit: 662

Quote from: joker_josue on June 01, 2024, 03:47:10 AM

I believe that the real differentiating point will be the coherence between texts. The AI will generate each text individually, and not as an individual person who has a pattern of feelings, tastes, attitudes, etc.

For example, in a post the IA may write that he loves eating apples. But 10 posts later he can already say that he hates apples and that he is allergic.

Even before AI become popular, there have been many users who contradict what they said or probably they have Dissociative Identity Disorder.

Those users will say people need to use decentralized exchange due to privacy and non custodial wallet to hold coins, but if you check their post history, they actually use Binance as their wallet. Not sure how should the forum react with those kind of users since they're not breaking any rules.

Topic: Lol - page 2. (Read 565 times)