7 ms·

"If you don’t like this level of deceit, you’re going to hate nearly all applications of Generative AI because they do far worse and on a far larger scale." No

"If you don’t like this level of deceit, you’re going to hate nearly all applications of Generative AI because they do far worse and on a far larger scale." Nope, if ChatGPT generates me some code that I can use for a real problem, or a generates product description for a friends buisness, than this is generating real value and no fake, no deceiving, no lie. What value for society is your buisness adding? "1) Pray tell, how did the forums grow from zero discussion to being a vibrant place, without some team of people “astroturfing” discussion?" By people sharing genuine interests. Not everything is a lie. But you are right, too much is already. So it is in no way ethical to add more lies to the pie.

EGreg 4y ago
If ChatGPT generates an essay for you, or a homework assignment, that IS fake, sorry. You didn’t do your homework. You didn’t write the essay. You’re passing it off like you did. If MidJourney “painted” your painting, and you pass it off as your own to others, you’re lying. This guy won a photography contest with a non-real photo: https://www.cbsnews.com/amp/news/artificial-intelligence-photo-competition-won-rejected-award-ai-boris-eldagsen-sony-world-photography-awards/ https://www.cbsnews.com/amp/news/artificial-intelligence-pho... He was honest enough to reveal it and reject the prize. But had he not done that, humans and honest photos would be out of the running just as if a chessplayee with a hiddenchess engine entered a tournament. It’s cheating, plain and simple. If you were honest, you’d reveal how you did the job and allow the people to choose whether to even have you in the loop, or continue to pay you as much. Even your product description example is fake - the AI didn’t experience the product, didn’t use it, can’t vouch for anything in the product. At least you are in the loop and spot-check the work for accuracy before submitting it. But an AI “describing” a product it has never seen is inherently as fake as a photograph of a scene that never happened. That is what generative AI does. It allows you to generate things without doing the work. I’m not talking about compilers of higher-level languages. I’m talking about literally not doing as much of the work as you want. Like the whole essay. You practice nothing. You cheat yourself. That in itself is bad but the much worse part is that people will be offered SERVICES to do it at scale. THAT IS THE PROMISE OF ALL APPLICATIONS OF GENERATIVE AI: to generate good looking results at scale! You claim to be appalled by a service that does what nearly every single forum owner already did with human workers. Why aren’t you grasping that this is the raison d’etre of the vast majority of AI applicatioms? You disingenuously listed a use case where you just logged into a UI an used it yourself, but obviously the APIs of OpenAI are offered to applications to do this at scale. Mine is just one of the most ethical ways to do it. I voluntarily put guardrails into it. But nearly EVERY OTHER USE OF THE API is of this nature but far worse. And yes I’d rather MY service get adoption than a far less scrupulous one. The value it is adding is generating discussion around topics that would otherwise receive no traction not becauae they aren’t interesting or the article not well researched and presented but because they don’t have “social capital” or a swarm of people fake-liking and fake-upvoting it. Here on HN and all other platforms those that have a few such people have an advantage. How do you think every single piece of content they post gets massive upvotes and likes? This levels the playing field for people who don’t have money to hire fake shillers and employ a team of humans to deceive others, in what is inherently a deceptive activity (astroturfing). Now everyone can have lots of interesting hooks for discussion, and the bots are clearly labeled as bots! Unlike many times the friends and employees of the owner don’t bother to disclose their relationship. It actually improves the situation in that respect vs the human shill team. That is just some of the value it gives, and then it will also soon answer helpdesk questions. I can bet you that you will be talking to a lot of bots soon for helpdesk questions, who won’t reveal they are bots — and somehow you won’t be shaming all those companies — it’ll become the norm, many people will be fooled into thinking that a real human took the time to emphathize with them and patiently help them, and they’ll treat the service nicely and pay them more according to that misunderstanding of “white glove service” rather than treating it as the cheap commodity that it is. In fact that’s at the heart of capitalism and the profit motive… when a new technology appears, or a consultant automates their own job, they keep quiet about it so they can still collect the profit as others think they are doing it all by hand. Employees who do this know they will be fired or they pay cut if they automated themselves out of a job. Capitalism breeds many inherently dishonest motivations, and also greed and jealousy of course. Having said that… you also totally ignored my request to answer point by point the numbered points. You seem to be arguing in bad faith since you are quick to shame others and do not seem to be actually resolve any misunderstandings or disagreements. Answer the points I raised in your next response, or it will be obvious to everyone that you’re dodging 99% of the substance in order to make a tortured point which is, by and large, completely the opposite of reality.
- hutzlibu 4y ago
  "Here on HN and all other platforms those that have a few such people have an advantage. How do you think every single piece of content they post gets massive upvotes and likes?" Yeah, this is the key point. Voting rings etc. exists also here, but they are against the rules and are spotted and banned regulary. So in general the pieces with the most upvotes are those where the majority of REAL people thinks they are interesting, like it should be. The same with genuine blogs, forums etc., those with INTERESTING content gets to the top. But if I encounter a blog with artificial engagement, I will just go away. "Having said that… you also totally ignored my request to answer point by point the numbered points." And why do you think I would feel obliged to do so, when you ignore my concrete criticism? You are creating lots of text defending the "ethics" of your buisness model, when in truth your buisness models depends on you keeping your twisted sense of ethics, no thank you to such a discussion. Because you are not discussing in good faith. "If ChatGPT generates an essay for you, or a homework assignment, that IS fake, sorry" Because I was not talking about homework, but real code I write. That code works in the real world. Real value. And the buisness description describes a real product and the content came from a real human, just the structuring of the words came with help. Framing this as cheating, so your own very shaddy things comes of as better, is just not something I will engage with any further.
  - EGreg 4y ago
    I know, you were trying very hard to list a tiny, tiny subset of activities people do with OpenAI’s APIs, so as to steer clear of the vast majority of use cases where you would have to concede the point. And you again avoided answering my points. At this point it’s rather clear you’re not arguing in good faith. Totally disagree that it is because of how good the content is. I posted the same exact thing twice - one time it got 0 or 1 upvotes and the other time it got 150 and made the front page! Same exact link. Rather, it is survivorship bias. You see the stuff thag made it. The celebrities on Twitter have entire departments working to make sure their content has likes and comments. In fact a decade ago I worked in digital agencies thag made Facebook apps and they all bought 50,000 non-organic “likes” to bootstrap the campaign / app. Nearly everyone who is successful does it. Practically any channel or forum that you’ve ever heard of got that way because they started off with some dedicated people / team members hanging out there every day and not behaving “organically”. Even worse, every very successful forum or celebrity got there by leveraging the existing social capital of some network or organization. Most of it is NOT organic at ALL. I am surprised that you would be so naive to think that putting up a forum with no comments is all you need to do to become a well-trafficked forum. PS: I brought up your concerns, which I share regarding the vast majority of AI applications, in a Discuss HN post. It got 14 upvotes and 7 comments in a few minutes, and then was flagged. Two hours later it was unflagged but by that time all momentum has cooled. I can assure you that far larger business models are at stake and far larger interests do not want this conversation to happen, than my little app. If you actually cared about this issue, you’d care about seeing it everywhere it exists: https://news.ycombinator.com/item?id=35688266 https://news.ycombinator.com/item?id=35688266

EGreg 4y ago

If ChatGPT generates an essay for you, or a homework assignment, that IS fake, sorry. You didn’t do your homework. You didn’t write the essay. You’re passing it off like you did. If MidJourney “painted” your painting, and you pass it off as your own to others, you’re lying. This guy won a photography contest with a non-real photo: https://www.cbsnews.com/amp/news/artificial-intelligence-photo-competition-won-rejected-award-ai-boris-eldagsen-sony-world-photography-awards/ https://www.cbsnews.com/amp/news/artificial-intelligence-pho... He was honest enough to reveal it and reject the prize. But had he not done that, humans and honest photos would be out of the running just as if a chessplayee with a hiddenchess engine entered a tournament. It’s cheating, plain and simple. If you were honest, you’d reveal how you did the job and allow the people to choose whether to even have you in the loop, or continue to pay you as much. Even your product description example is fake - the AI didn’t experience the product, didn’t use it, can’t vouch for anything in the product. At least you are in the loop and spot-check the work for accuracy before submitting it. But an AI “describing” a product it has never seen is inherently as fake as a photograph of a scene that never happened. That is what generative AI does. It allows you to generate things without doing the work. I’m not talking about compilers of higher-level languages. I’m talking about literally not doing as much of the work as you want. Like the whole essay. You practice nothing. You cheat yourself. That in itself is bad but the much worse part is that people will be offered SERVICES to do it at scale. THAT IS THE PROMISE OF ALL APPLICATIONS OF GENERATIVE AI: to generate good looking results at scale! You claim to be appalled by a service that does what nearly every single forum owner already did with human workers. Why aren’t you grasping that this is the raison d’etre of the vast majority of AI applicatioms? You disingenuously listed a use case where you just logged into a UI an used it yourself, but obviously the APIs of OpenAI are offered to applications to do this at scale. Mine is just one of the most ethical ways to do it. I voluntarily put guardrails into it. But nearly EVERY OTHER USE OF THE API is of this nature but far worse. And yes I’d rather MY service get adoption than a far less scrupulous one. The value it is adding is generating discussion around topics that would otherwise receive no traction not becauae they aren’t interesting or the article not well researched and presented but because they don’t have “social capital” or a swarm of people fake-liking and fake-upvoting it. Here on HN and all other platforms those that have a few such people have an advantage. How do you think every single piece of content they post gets massive upvotes and likes? This levels the playing field for people who don’t have money to hire fake shillers and employ a team of humans to deceive others, in what is inherently a deceptive activity (astroturfing). Now everyone can have lots of interesting hooks for discussion, and the bots are clearly labeled as bots! Unlike many times the friends and employees of the owner don’t bother to disclose their relationship. It actually improves the situation in that respect vs the human shill team. That is just some of the value it gives, and then it will also soon answer helpdesk questions. I can bet you that you will be talking to a lot of bots soon for helpdesk questions, who won’t reveal they are bots — and somehow you won’t be shaming all those companies — it’ll become the norm, many people will be fooled into thinking that a real human took the time to emphathize with them and patiently help them, and they’ll treat the service nicely and pay them more according to that misunderstanding of “white glove service” rather than treating it as the cheap commodity that it is. In fact that’s at the heart of capitalism and the profit motive… when a new technology appears, or a consultant automates their own job, they keep quiet about it so they can still collect the profit as others think they are doing it all by hand. Employees who do this know they will be fired or they pay cut if they automated themselves out of a job. Capitalism breeds many inherently dishonest motivations, and also greed and jealousy of course. Having said that… you also totally ignored my request to answer point by point the numbered points. You seem to be arguing in bad faith since you are quick to shame others and do not seem to be actually resolve any misunderstandings or disagreements. Answer the points I raised in your next response, or it will be obvious to everyone that you’re dodging 99% of the substance in order to make a tortured point which is, by and large, completely the opposite of reality.

hutzlibu 4y ago
"Here on HN and all other platforms those that have a few such people have an advantage. How do you think every single piece of content they post gets massive upvotes and likes?" Yeah, this is the key point. Voting rings etc. exists also here, but they are against the rules and are spotted and banned regulary. So in general the pieces with the most upvotes are those where the majority of REAL people thinks they are interesting, like it should be. The same with genuine blogs, forums etc., those with INTERESTING content gets to the top. But if I encounter a blog with artificial engagement, I will just go away. "Having said that… you also totally ignored my request to answer point by point the numbered points." And why do you think I would feel obliged to do so, when you ignore my concrete criticism? You are creating lots of text defending the "ethics" of your buisness model, when in truth your buisness models depends on you keeping your twisted sense of ethics, no thank you to such a discussion. Because you are not discussing in good faith. "If ChatGPT generates an essay for you, or a homework assignment, that IS fake, sorry" Because I was not talking about homework, but real code I write. That code works in the real world. Real value. And the buisness description describes a real product and the content came from a real human, just the structuring of the words came with help. Framing this as cheating, so your own very shaddy things comes of as better, is just not something I will engage with any further.
- EGreg 4y ago
  I know, you were trying very hard to list a tiny, tiny subset of activities people do with OpenAI’s APIs, so as to steer clear of the vast majority of use cases where you would have to concede the point. And you again avoided answering my points. At this point it’s rather clear you’re not arguing in good faith. Totally disagree that it is because of how good the content is. I posted the same exact thing twice - one time it got 0 or 1 upvotes and the other time it got 150 and made the front page! Same exact link. Rather, it is survivorship bias. You see the stuff thag made it. The celebrities on Twitter have entire departments working to make sure their content has likes and comments. In fact a decade ago I worked in digital agencies thag made Facebook apps and they all bought 50,000 non-organic “likes” to bootstrap the campaign / app. Nearly everyone who is successful does it. Practically any channel or forum that you’ve ever heard of got that way because they started off with some dedicated people / team members hanging out there every day and not behaving “organically”. Even worse, every very successful forum or celebrity got there by leveraging the existing social capital of some network or organization. Most of it is NOT organic at ALL. I am surprised that you would be so naive to think that putting up a forum with no comments is all you need to do to become a well-trafficked forum. PS: I brought up your concerns, which I share regarding the vast majority of AI applications, in a Discuss HN post. It got 14 upvotes and 7 comments in a few minutes, and then was flagged. Two hours later it was unflagged but by that time all momentum has cooled. I can assure you that far larger business models are at stake and far larger interests do not want this conversation to happen, than my little app. If you actually cared about this issue, you’d care about seeing it everywhere it exists: https://news.ycombinator.com/item?id=35688266 https://news.ycombinator.com/item?id=35688266