Topic: bitcointalk forum vs OpenAI - ChatGPT - page 9. (Read 3190 times)

The first article (semi-) written by chatGPT is here: https://notatether.com/servers/everything-you-need-to-know-about-servers/

To be honest, only the paragraphs are AI-generated. The prompts I gave it were adapted to headings, and I added some disclaimers and an affiliate paragraph and the introduction, and some other sentences, by hand.  I found that the AI-generated images from DALL-E were disastrously off-topic and irrelevant for the subject matter, so all images are just sourced from free sites.

Writing the prompts and copying the answers took less than 10 minutes. We'll see how it does with SEO. I plan to make more articles like this to see how it affects my site in general (after all, JARVIS who runs the Iron Legion is writing them! )

PS. Don't worry, this thread won't become an unsolicited advertisement for my site.

Fuck it, I am relaxed now LOL
I thought it will bring up my personal details (unlikely) 🤣
Well to be honest I was expecting it to bring some forum posts that I made and things like that because that's what they will get if they want to connect anything with the handle. It does not even have any clue of my original username.
Eventually I realize there are no intelligence. The program was written in some way to create words and sentences and make them meaningful before showing it to you. Additionally, they are scrapping texts from all websites the same as a search engine does and giving your pointless information. However, as a human you get a clue of the topic you are searching and able to create your own writing.

Money are the data they are taking from you. Not everyone is creating account with throwaway details even if this is a email address.

I meant paying for AI art under false premise. I think it would be fraud to sell such art as being made by "a famous artist" just like it's fraud to sell a fake Monet. Even if a double blind study can't tell the difference.

I say I'm not going to do it for lack of time mainly, but if someone did, I don't think it would be very smart to let it show with the name that an AI is being used, as I think the forum members would be more reluctant at first to give merit to the account.

Name the account OpenAI - ChatGPT

---

---

Just a thought: have you realized you're not using the free chat AI, you're training it?

Instead of the negative, I think the leverage opportunity is here.

I don't have time for more, but I would like to try to create an account that would write posts exclusively with the help of AI, revised and rewritten adapting them to each case, obviously. And with citations to avoid plagiarism.

I think the result would be to do more with less. More quality in less time.


Sounds like science to me:

https://www.cancer.gov/publications/dictionaries/cancer-terms/def/double-blind-study

Probably. It doesn't "know" things, it gets it's information from other sources. But it also gives the wrong information, see how quickly Stack Overflow banned them. Dutch media had an article about it today, they described it as "AI makes predictions". I think that's a good way to describe it: it's a model, and that model can be improved by feeding it more data. But it can easily be wrong, and without manual review can give incorrect answers.

I am still little confuse about this AI bot.  When you say that this bot will combine the answers with different sources and will come up with the best possible answer, does it mean it will copy the exact same lines from different sources and produce plagiarised content ?

There are other intelligent bots available in the market (Good ones are paid), which will give you plagiarism free articles for your sites etc. So i wonder if the OpenAI - ChatGPT free software is comparable to those paid ones 

So just s little update, I'm finding the workflow of just pasting stuff from chatGPT much faster. I still have to write the headings obviously, but I was thinking if I combine it with an AI art generator, that should fulfill the image requirements a normal article would have.

I'm excited to see what the results are once I start mass-producing these (for now restricted to one category). What used to take several days now takes only a few mins max (except for the images part, which I'm still working on)

nope. None .
14million startup
About 180k a day right now .
Organized by a nonprofit .


Just imagine how other industries have implicated it uses.
If you don't use social media aside from Bitcointalk it's likely beyond your comprehension of the impact it's already had.
Good thing these systems have been internal!
 Imagine if the non profit groups have accomplished this much , imagine how good the for profit groups have been doing the last 8 years or so.

This linguistics model has already changed the world .   You won't start seeing results untill q1/q2  and that's only because it was publicly launched .

This shits already gated fyi. The little guy thinks he has a chance writiting white papers  and articles, business plans  code ECT .... He doesn't . ( A few beers in on a rant sorry haha)

This is likely what will happen. I mean, the whole generating art via AI has moral implications which haven't been really highlighted, and discussed about yet. However, you also bring up a good point. A lot of people are generally quite lazy, and will look for the easiest way to make money, sites which get people to write their articles, and then pay them based on views will likely be inundated with this type of content, which will ultimately lead to the major search engines delisting anything that's found to have this type of content. 

Unfortunately, this could be rather problematic for forums, and discussion websites where freedom of speech is encouraged. We could moderate this sort of content, but if it the technology is likely to develop even further, and the current state of it being quite generic, and subtly able to tell the difference, that margin could lesser in the near future.

I don't know whether that's true, since I imagine the software has a ton of data to read from, and generate its content. So, the data source they're drawing from probably is already had it's max efficiency, and more data doesn't necessarily mean it'll get better. Developers implemented ideas could, but I imagine before this really takes off, it will already be causing companies, and search engines a ton of problems.

They could paywall it, but ultimately people would be willing to pay for this type of service, since it's very likely they'd be generating content, and earning from it.

I don't think they're making any money off of this service, it's not like the rest of the OpenAI services where you have to pay money to get API credits. I haven't seen anything like that on the site otherwise they would have been marketing those figures by now.

As per CEO of OpenAI, it took just 5 days to cross 1 million users.


https://twitter.com/sama/status/1599668808285028353

Sounds like fraud.


Ok, I would agree with that, some low-level tasks is where the AI could shine with its wikipedia skills.



To me that just means that more likely the whole "genre" will die off, rather than produce some sort of real value. If the barrier of entry is too low and AI texts flood it, then traffic will decrease to the point of unprofitable or Google et al will demonetize it.

I want to check this out for myself. Maybe I will make a section on my site just for AI posts called JARVIS or something like that, and report here how much traffic I get from that on Google.

Of course it will not write about anything technical

If they don't know it's made by an AI and are convinced that a famous artist has made them, I'm pretty sure they will do. I'd like to see double blind experiments with this.

Besides, the question is not so much how much can be paid, but how much can be saved by companies that currently have draftsmen on staff.

Going back to OpenAI, at the moment it's capped and doesn't get real time information, but I imagine that in the future that will be the case. Does anyone know if this is the case?

As I commented in the other thread, I already saw some time ago that there are blogs that generate revenue written in good part by AI.

DALL-E is cool but also more in an amusing than practical way. I've seen some experiments on Twitter feeding ChatGPT generated texts into MidJourney and the results are quite interesting but I still struggle to see a practical application. Is someone going to pay money for "art" generated by AI?

Yes maybe, but not today



You just need to ask politely "citation please"

or ask just "write a post about bitcoin technology but please don't plagiarize the text"

---

Let the AI answer this question too.

(You can jump to the last paragraph)



Certainly it's not going to replace humans in any way. But it is going to change many things, I don't know yet but I can feel it... You must have seen those DALL-E AI-generated images like this, these are coming from same OpenAI.

Combine these ChatGPT Text + DALL-E images, you are ready to take over humans (on the internet) in terms of speed and some accuracy too.

---

I did try this too but the answer was same like in the case of LoyceV, "BitcoinGirl.Club could potentially be a website, a username or handle used by someone on the internet, or even the name of a person or group."

Oh I just noticed the answer to your question "Who is BitcoinGirl.Club?" is different from the answer to "Who's LoyceV?"



Based on your username it perfectly deduced
1. That it could be a website
2. It could be a group  
3. It could be a username

If we come across a positive and politely worded post, we know it's ai🤣. Please don't ask ai for advice.

Me getting ready to start merit farming after reading this topic, lol.

For the forum? Not really. I don't know about you but I don't come here to read generic vapid faceless texts, I come here for a discussion, to learn stuff that I can't easily google, maybe for an occasional fight or trollbait. I can imagine that due to the monetization peculiarities of this forum AI-generated texts could drown out useful/interesting/entertaining content even more than current spammers do, and either kill the forum completely or force theymos to ban signatures/avatars/etc.


Again, "biggest" for what? Bitcoin solves a problem. AI chat seems like a fidget spinner. Nice to play with for a few minutes. Useless for reliable information (better off going to e.g. the aforementioned stack overflow directly).


Ok, credit where credit's due - it didn't attempt to kill you. But didn't really provide any useful information for fixing the issue either. Safe, useless, and probably what a public-use AI will have to be lest it be sued out of existence.