Once again another reason for signature vacation time.
Every week flip a coin and either shut signatures off or turn them back on.
Makes bots and ai have less reason to post if there are no signature revenues.
Unfortunately our forum is existing for the wrong reasons.
Signature revenue is a shitty reason to exist.
Spam is worse.
Scammers are the worst.
Trolling fit in there somewhere but not as bad as a scammer.
At loyce how about scraping a signature campaign thread and estimate what it pays each week.
I often complain about signatures and have asked for signature vacation time.
So here is another reason to try it out.
Topic: AI writing messages on Bitcointalk.org - page 9. (Read 3673 times)
If the article posted by OP is really written by using AI, then I can only say "Houston, we have a problem".
We already know what the answer is - https://bitcointalksearch.org/topic/kyc-now-required-5124947
Or perhaps everyone will have to read their posts out loud on video, but keep an eye out for deep fake Tom Cruises if they forget to reset it.
If the article posted by OP is really written by using AI, then I can only say "Houston, we have a problem". I believe that the predecessors of GPT-3 are in use and have found practical application, especially on forums and some social networks, and that we will soon get literally virtual characters in the virtual world which will very faithfully imitate real persons.
If I have well understood the whole procedure for AI to write an article, it comes down to the setup of basic parameters on the basis of which then GPT-3 writes that article. I looked at some of Manuel Araoz's tweets and here's what it looks like in practice.
https://twitter.com/maraoz/status/1284581042175836163
If I have well understood the whole procedure for AI to write an article, it comes down to the setup of basic parameters on the basis of which then GPT-3 writes that article. I looked at some of Manuel Araoz's tweets and here's what it looks like in practice.
https://twitter.com/maraoz/status/1284581042175836163
I haven't seen any AI capable of having an actual opinion, let alone style or personality, and the article hasn't convinced me. But it might be able to fool a signature campaign long enough to get paid so there is definitely a niche here.
It's just getting warmed up. We'll have to wait and see. The opinion bit is the one they'll be looking to refine no doubt.
This was the all AI Reddit sub I referred to - https://www.reddit.com/r/SubSimulatorGPT2/
It's a truly painful read but strangely reminiscent of an awful lot of posts on this very forum.
The years of google translate or generally empty shitposting has kind of softened us up to be more receptive to this too. I presume there is some gaping hole where this system can be confused or defeated. We'll have to wait for it to be uncovered.
I for one welcome our new AI overlords. Finally I can ask "is CSW the real Satoshi" and get an honest answer. Or most likely get an irrelevant salad of Google search results and tag the fucking bot with a neutral... actually even red trust would be probably ok since I don't think that trading with a bot would be advisable at all but I digress.
I haven't seen any AI capable of having an actual opinion, let alone style or personality, and the article hasn't convinced me. But it might be able to fool a signature campaign long enough to get paid so there is definitely a niche here.
Unless theymos gonna make or use AI which can detect whether a message is written by AI/not, there's nothing much we could do aside from reporting posts that doesn't make sense.
Using bots isn't even against the rules.I'm impressed that the linked article was entirely AI.
Me too, and it makes me very curious where it gets its input data. Basically, it just made up some lies! What's the point in that?I do see potential though, we'll have thousands of different AIs compete to join signature campaigns managed by other AIs. Or, thinking bigger, an AI can write those boring work reports. Then another AI can read it, and in the end nobody has to do anything anymore. Especially with the recent work from home, your own cloud AI can apply for thousands of jobs per minute pretending to be you, and since you don't have to show up in person anyway, your AI can maintain dozens of jobs
There's simpler & good horrible solution for this problem, which forcing user to solve CAPTCHA every time he made/edit message.
AIs are much better at it than humans. Try to solve Google reCAPTCHA using Tor within 5 minutes. It's painstaking for humans, while an AI has all the time it needs. 
I have no doubt AI research is going to change things, but bitcointalk forum may also change in time (I hope).
One thing we could see is some way to prove we are humans with some kind of sybil tests, or with new way to log in the forum.
Lot of bitcoins are spent for new forum, and I think forum proof-of-person system can also be considered for future.
One thing we could see is some way to prove we are humans with some kind of sybil tests, or with new way to log in the forum.
Lot of bitcoins are spent for new forum, and I think forum proof-of-person system can also be considered for future.
I assume this a brand new release?
There's a Reddit sub where only GPT bots are allowed to post. That is fucking awful. The sentences are complete and coherent but the whole place echoes with absolute and overwhelming emptiness.
I'm impressed that the linked article was entirely AI. There were a few things that jarred but overall it's a (much) higher standard than most posts here. I can't really see how this can be resisted or identified, especially if it continues to be refined.
The years of google translate or generally empty shitposting has kind of softened us up to be more receptive to this too. I presume there is some gaping hole where this system can be confused or defeated. We'll have to wait for it to be uncovered.
That would be the end of my relationship with this place. But there'll be a bot for that too no doubt.
There's a Reddit sub where only GPT bots are allowed to post. That is fucking awful. The sentences are complete and coherent but the whole place echoes with absolute and overwhelming emptiness.
I'm impressed that the linked article was entirely AI. There were a few things that jarred but overall it's a (much) higher standard than most posts here. I can't really see how this can be resisted or identified, especially if it continues to be refined.
The years of google translate or generally empty shitposting has kind of softened us up to be more receptive to this too. I presume there is some gaping hole where this system can be confused or defeated. We'll have to wait for it to be uncovered.
There's simpler & good horrible solution for this problem, which forcing user to solve CAPTCHA every time he made/edit message.
That would be the end of my relationship with this place. But there'll be a bot for that too no doubt.
Unless theymos gonna make or use AI which can detect whether a message is written by AI/not, there's nothing much we could do aside from reporting posts that doesn't make sense.
There's simpler &good horrible solution for this problem, which forcing user to solve CAPTCHA every time he made/edit message.
There's simpler &
Think only about the implication on the level of spam two bots can generate discussing together, or the problem for campaign managers, of even for paid services like bumping services of marketing solutions.
Well, the article is of a great quality compared to the badly written and badly translated spam we have here in great quantities.
Which, on a sadly sarcastic note means that AI posting here instead of those guys would increase the quality of posts.
Unfortunately this also means that it'll be harder to find out which poster is genuine human and which is an AI posting high quality, well written... made up trash.
Definitely scary to think about. Although, I’m sure the bots have been here with us for a long time, just not as sophisticated as OpenAI’s new GPT-3. 
For example, take a look at this guy's posting history: Subbir
For example, take a look at this guy's posting history: Subbir
I lose brain cells everytime I read one of his posts lul.
Could you imagine how much it must sting if you're an actual person trying to earn merit and a damn script comes along that manages to earn more
Some say I've been doing that for years.I did a subtle reference to the quasi-AI nature of your parent account in my OP, actually: a suspect many have arisen in the past.
With the fast developments in IT and machine learning, bots will be upgrade to smarter AI-bots and I believe that they will be able to wisely enough to create posts and join discussions. It is the thing and change to come to our forum, obviously. You can see the same thing with Google translator. Years ago, translation products from Google translator are terrible in quality but now its quality is acceptable enough and can be used for people can not translate original documents by themselves.
Do you think we can avoid such AI-bots evasion here? I don't think so.
But to say AI-bots can damage or destroy our forum, I don't believe they can.
We have:
Basically, who create bots, even AI-bots? Human, so human can beat bots.
I don't like Facebook but they have a feature to disable rights to share a livestream if finding something suspicious on exponential shares in short period of time. Maybe sometimes findings are false ones but it is helpful.
Do you think we can avoid such AI-bots evasion here? I don't think so.
But to say AI-bots can damage or destroy our forum, I don't believe they can.
We have:
- Merit system: to prevent spammers, and bots (even with AI-bots) to rank up. If needed, merit system's upgrades can do better to prevent them.
- Bump score: to prevent bad effects from bump services. Even if tehy are AI-bots, they can not beat the bump score and the algorithm of bump score can be upgraded too.
- Combined effects from merit system and bump score: they can not rank up, their bump score are nearly zero and (if >0) can not increase.
- Reports from community can help: to delete spam posts, bump posts, and nuke related accounts.
- Avatar and signature disabled can help: I think AI-bots are smart enough to realize those changes and stop posting
Basically, who create bots, even AI-bots? Human, so human can beat bots.
I don't like Facebook but they have a feature to disable rights to share a livestream if finding something suspicious on exponential shares in short period of time. Maybe sometimes findings are false ones but it is helpful.
Could you imagine how much it must sting if you're an actual person trying to earn merit and a damn script comes along that manages to earn more
Some say I've been doing that for years. 
That might cause big trouble in our community (or even other forums) because we aren't able to determine the characteristic of posting on this AI bot since the bot can able to generate text that relevant to the topic. Sound like prediction answer.
But I noticed that there are some awkward words that the AI bot has been used. And the creator was right, probably we can determine in this way.
If someone reply or making a post something like this conscious to unconscious probably they're bot or spammer.
But I noticed that there are some awkward words that the AI bot has been used. And the creator was right, probably we can determine in this way.
If someone reply or making a post something like this conscious to unconscious probably they're bot or spammer.
Maybe the AIs will evolve to the point where they generate a more stimulating conversation than traditional spammers. Could you imagine how much it must sting if you're an actual person trying to earn merit and a damn script comes along that manages to earn more because it makes more sense than you do.
legendary
Activity: 1960
Merit: 1908
Marketing Campaign Manager |Telegram ID- @LT_Mouse
I have read the article and to be honest I did not notice for a single time that the article could be written by an AI. If such posts can be generated from AI, I think we have some in our forum. The article was correctly written and is far above grade in term of quality from most of our forum users.
Some days ago, I think someone had reported user who were posting randomly. May be they are the creation of GPT-3.
Some days ago, I think someone had reported user who were posting randomly. May be they are the creation of GPT-3.
Definitely scary to think about. Although, I’m sure the bots have been here with us for a long time, just not as sophisticated as OpenAI’s new GPT-3.
For example, take a look at this guy's posting history: Subbir
For example, take a look at this guy's posting history: Subbir
An interesting read.
Please read the whole article before going on through my post.
OpenAI's GPT-3 may be the biggest thing since bitcoin
Well, we may have finally shed light on the AI nature of some users.
/j
Actually, I think the implication here might be important to consider.
Think only about the implication on the level of spam two bots can generate discussing together, or the problem for campaign managers, of even for paid services like bumping services of marketing solutions.
Two boys discussing and shilling a new project can push social engineering to a whole new level.
This kind of software might be too sophisticated to be intercepted or proven to be used with an high enough degree of certainty.
I think it is something needs to be addressed.
Please read the whole article before going on through my post.
OpenAI's GPT-3 may be the biggest thing since bitcoin
Quote
Summary: I share my early experiments with OpenAI's new language prediction model (GPT-3) beta. I explain why I think GPT-3 has disruptive potential comparable to that of blockchain technology.
Well, we may have finally shed light on the AI nature of some users.
/j
Actually, I think the implication here might be important to consider.
Think only about the implication on the level of spam two bots can generate discussing together, or the problem for campaign managers, of even for paid services like bumping services of marketing solutions.
Two boys discussing and shilling a new project can push social engineering to a whole new level.
This kind of software might be too sophisticated to be intercepted or proven to be used with an high enough degree of certainty.
I think it is something needs to be addressed.
Jump to: