Spotify being listed as one of the brands doing AI right is jarring to me. It’s the only one of the companies mentioned where I’ve had multiple conversations with non tech people about how much they hate their AI usage. Specifically in their case the drift to AI produced and performed music in the playlists that are not explicitly labeled as AI. And there seems to be a nuance to what uses people are ok with, large language models for personal use, ok, generative ai in any creative capacity, offensive. Obviously anecdotal, but this article was very far from the reality I’m experiencing
Now, personally I want AI to disappear, but music is one of the few areas where AI COULD be potentially useful. The reason I say this is because for music, the criterium should be whether the music is good or bad, not who made it. I still think humans are better than AI, but there were audio tracks fully AI generated which was not bad. It does not convince me into one of those humans who fell for the skynet trap and embraced AI, but I also can not say that all of AI is totally useless - just 99.9%.
> the criterium should be whether the music is good or bad, not who made it.
Setting aside how simplistic a simple good/bad scale for any art is, who made a piece of music and what they were expressing with it is absolutely part of the quality consideration.
I feel that "good" music has some emotional resonance with me, at some level. In my opinion AI won't ever be able generate what I'm looking for, it always strikes me as somewhat bland or seems too much like another track.
Background music in distracting environments, where the listening is maybe the third or fourth thing I'm doing, maybe AI will be acceptable.
https://www.audible.ca/fr_CA/pd/What-Could-Go-Wrong-Livre-Au... , this podcast (which I guess is now only available as a book) was really interesting, about 3/4 of the way through they tackle the emotional resonance bit, with one of the people from Broken Social Scene, and the artist kinda breaks down because the AI model they were experimenting with was able to replicate the "humanness" of music that loved. All in all it was a really interesting program, but the music section was probably the most fascinating.
The humanness of the music I listen to is given by actual humans composing and playing it. I don't mean little mistakes or whatever signal, but knowing it's made by humans. Otherwise I can't be arsed to listen to it, because probably I only care so much about the music itself. Because why would I? But yeah some people choose a virtual girlfriend too, so I'm probably just not in the right target group for AI productions.
PS: of course they can lie to me. Until I find out.
> The reason I say this is because for music, the criterium should be whether the music is good or bad, not who made it
The first thing I do when I find a nice track is to lookup the album and the artist. I may not like all the tracks on the album, or all the albums of the artists. But the album vision and the artist motivation and skills is why an album (I listen by album) get added to my collection. That collection is meaningful for me, not for the quality of the tracks (which is fairly subjective) but the emotional resonance I have with it. Kinda like the house I grew up vs a random penthouse. An AI produce piece may be nice, but it won’t have the emotional depth to anchor it for me.
But then why stop at music and not apply it to all art like movies and books as well. The criterium to judge should be the same for them.
But I feel you are underestimating how much people care about the artist as well. Could AI get that level of support from people knowing its not a real human being? I don't think so. I mean most artist stories are also fabricated but if you explictly tell people hey this is like not real at all and fully ai but its a rapper from this particular city who grew up struggling with this particular issue so people from that city and facing those issues could relate to him probably, dont think it will work because people will still know its a bot pretending to be from their city and facing their particular issue.
Lastly i also generally public likes authenticity and the idea of an algorithm just spamming music is not very inspiring in that regard. Maybe in a blind test, they wouldn't even be able to tell which track is ai and which is not but the moment you tell them which one it is, they will likely think its bad, because who made the art completely changes the prespective of the consumer
> Spotify's personalization exemplifies invisible AI done right.
Ironically, I can't think of a platform that does personalization worse. Not only does it regularly surface music I don't like, it surfaces the exact same music I don't like over and over again. I don't know whether this is bad recommendation or their pushing music on me (ie payola) but it's supremely irritating.
I would have thought that adding semantic search to a photos app might be a better example of good AI, not these bolted-on-top examples.
Yep. And if you want to integrate AI into your products, like we are currently doing, you are best off just delivering features / products and concealing the fact it's "AI" instead of putting it front and centre. AI has a rather tarnished brand amongst many groups of people.
I dislike AI because the result has consistently been bad. The most enthusiastic AI co-workers are producing garbage at a 100x rate, while the non-enthusiastic responsible ones are left reading and reviewing it.
More subtle thought-out AI tends to work better than AI for the sake of it anyway. If you deliver something useful, people will end up using it. A lot of current AI use is not particularly useful though.
Remember to flag it, I've been doing this consistently with AI slopicles. Eventually we can build a social consensus around this and stop having to see them.
It's kind of ironic, if this article was written about programming, it would be immediately flagged as being too thin and written almost completely with AI.
It's not just AI first companies..there's a place near me that tried to usr AI images in their social media ads. Completely blew up their relationship with the community. People would have preferred a cell phonr picture of their actual food to an AI generated approximation..
I've noticed this happening more and more! Mainly with restaurants! There's an indian place just down the street from my place with a menu composed of entirely (badly) AI generated curry pictures in the window. It turns everyone off eating there because it feels like they're somehow lying about what they sell. I get that curry is probably hard to photograph unprofessionally and still make it look appetizing, but I would've preferred
"Consumers have developed pattern recognition for AI-generated content."
The irony of reading an article that talks about AI slop that clearly seems to have been written by AI. Hey, I could be completely wrong, and it wasn't, but there are so many flags.
Do I care? Not really, but whoever wrote this is right. I guess we developed a pattern recognition for these things
Agreed. But it's also indicating the current crop of AI generates certain patterns that are detectable.
What if it advances to a point where people can't? Chances are that's anywhere between "we're already past that point" and "could be a decade or more away" depending on the type of work.
Imho AI is ultimately just a tool. Use where it improves things, avoid where it doesn't. Right now companies seem to try & shove it everywhere. Which clearly isn't working.
It almost feels like AI is a hype and is used by the NFT/blockchain/crypto/web3 grifters as yet another vehicle to grift people right into their graves. Fun times.
Always? Maybe for the past 15-20 years but for the 50 years before that there wasn't much investment or interest from executives. AI was mostly about research and niche applications.
AI has always been a marketing term. As soon as it has a useful application it is named around its utility. Search, recognition, generation, classification, etc. AI is the term you use when you can't easily describe what you're doing.
It's true that in the early years, much of the investment was public money—but the above still stands.
This is true and is the reason I personally have always hated the term. This time around things are different. This time we also hate the technology.
Personally I used to love Markov Chains, Kalman filters, Supervised learning, etc. and resent the fact that people called it AI. But LLM has nothing but hate from my heart. Ironically, I love that people call it AI, and do so enthusiastically my self. A bad and hyped up technology deserves a bad and hyped up name.
"When consumers believe emotional marketing communications are written by AI rather than humans, they judge them as less authentic, feel moral disgust and show weaker engagement and purchase intentions. This happens even when the content is otherwise identical."
Well, in general I do not care either way. I regard all ads as propaganda that attempts to steal my time. However had, even then it is indeed true that AI just is an additional annoyance factor, because it means that no real human really invested time - just AI slop that is spammed down onto people, and wastes their time. So I don't agree with the premise in the article to begin with, but most assuredly it is also true that AI slop just is pissing off people. I am noticing this on youtube too and although I don't have data, it seems that enough people were annoyed that the no-AI movement gained more grounds in the last some weeks. Hopfully we'll eventually reach AI extinction - not likely to happen, since some humans are already addicted to AI (see all "contributed with claude" on github spam), but I regard this as a noble goal. Rid this world of AI.
> Soon - and even now - there are people getting away with using it that you guys can’t detect.
Sure, but that seems largely irrelevant to what's being discussed here.
The public in general is anti-AI, so any marketeer thinking that associating AI with their brand is a positive is misreading the room, and any thinking that the public are unable to detect it most of the time, especially in something like a video, are just wrong.
Is that going to change - will AI-generated content become harder to detect? Sure, but "we know the public hate this, but we can con them" doesn't really sound like a recipe for success. In a world increasingly awash with AI generated content, it's easy to predict that human-generated content and interaction is going to be increasingly valued, and smart companies will realize this.
> Do HN luddites realize the technology is here to stay?
> Soon - and even now - there are people getting away with using it that you guys can’t detect.
It's very strange to me that this is the pro-AI argument. It's not that things will be better, its that they can't be fought effectively.
The inventors of the printing press and the steam engine and the internet--whatever parallel you want--had a vision of improvement, making things better. AI proponents seem to want to make things worse, gleefully, but not be criticised for it.
Even if you're completely correct, that's not a good starting point.
Strong disagreement on this one, I think the coverage has proven that the current crop if anti-LLM sentiment is a much bigger thing then whatever is happening on this website.
In the case of AI generated imagery, maybe it will get to the point where people can't tell just by looking at it. If it does, at what cost? It seems to me that token cost is rising, I can't help but wonder if hiring an illustrator might end up being less expensive in the long run.
Consumers and gamers are already alienated. If the price of RAM keeps going up I'm not sure much technology will actually be here to stay. What's the point of AI "art" anyway if everyone needs $10,000 devices to interact with the content? AI isn't going to raise wages, it's just going to enable people to produce more for the same wage.
Do HN luddites realize the technology is here to stay?
Oh, I get that. LLMs are here to stay, like pyramid and Ponzi schemes, like Bitcoin and NFTs. It will be here until it isn't or we aren't. (And it just may help with that last.)
LLMs are here to stay and I'm using them daily out of necessity and out of convenience too. I'm paying money for it too. It is a useful tech. Still, I hate it in general and the sociopaths promoting it in particular, and I foresee a great deal of issues for humanity because of that. Not catastrophic, just boringly bad. Like many humans going broke and jobless (booooring), or humans losing entire activities they actually liked to do to robots (booooring). It's sarcasm, if you didn't get it btw. I feel like I need to spell out a lot of things explicitly nowadays.
Like for example a lot of legitimate and useful companies are completely dependent on the advertisement, which is a net negative absolutely vile and disgusting invention. But we all collectively picked option to defect, and not because we are bad humans, and are paying for that decision collectively too.
Spotify being listed as one of the brands doing AI right is jarring to me. It’s the only one of the companies mentioned where I’ve had multiple conversations with non tech people about how much they hate their AI usage. Specifically in their case the drift to AI produced and performed music in the playlists that are not explicitly labeled as AI. And there seems to be a nuance to what uses people are ok with, large language models for personal use, ok, generative ai in any creative capacity, offensive. Obviously anecdotal, but this article was very far from the reality I’m experiencing
Same with the 'doing AI right' Netflix example for me. Showing me the same exact recommendations in each tray but in a different order isn't helping.
That is how netflix was before ai too.
Now, personally I want AI to disappear, but music is one of the few areas where AI COULD be potentially useful. The reason I say this is because for music, the criterium should be whether the music is good or bad, not who made it. I still think humans are better than AI, but there were audio tracks fully AI generated which was not bad. It does not convince me into one of those humans who fell for the skynet trap and embraced AI, but I also can not say that all of AI is totally useless - just 99.9%.
> the criterium should be whether the music is good or bad, not who made it.
Setting aside how simplistic a simple good/bad scale for any art is, who made a piece of music and what they were expressing with it is absolutely part of the quality consideration.
I feel that "good" music has some emotional resonance with me, at some level. In my opinion AI won't ever be able generate what I'm looking for, it always strikes me as somewhat bland or seems too much like another track.
Background music in distracting environments, where the listening is maybe the third or fourth thing I'm doing, maybe AI will be acceptable.
https://www.audible.ca/fr_CA/pd/What-Could-Go-Wrong-Livre-Au... , this podcast (which I guess is now only available as a book) was really interesting, about 3/4 of the way through they tackle the emotional resonance bit, with one of the people from Broken Social Scene, and the artist kinda breaks down because the AI model they were experimenting with was able to replicate the "humanness" of music that loved. All in all it was a really interesting program, but the music section was probably the most fascinating.
The humanness of the music I listen to is given by actual humans composing and playing it. I don't mean little mistakes or whatever signal, but knowing it's made by humans. Otherwise I can't be arsed to listen to it, because probably I only care so much about the music itself. Because why would I? But yeah some people choose a virtual girlfriend too, so I'm probably just not in the right target group for AI productions.
PS: of course they can lie to me. Until I find out.
I just can't see myself ever going to a venue to see an AI artist perform. Would it just involve watching a video?
How is this different from any other use of AI? Of course the criteria for everything should be good or bad not who made it.
> The reason I say this is because for music, the criterium should be whether the music is good or bad, not who made it
The first thing I do when I find a nice track is to lookup the album and the artist. I may not like all the tracks on the album, or all the albums of the artists. But the album vision and the artist motivation and skills is why an album (I listen by album) get added to my collection. That collection is meaningful for me, not for the quality of the tracks (which is fairly subjective) but the emotional resonance I have with it. Kinda like the house I grew up vs a random penthouse. An AI produce piece may be nice, but it won’t have the emotional depth to anchor it for me.
But then why stop at music and not apply it to all art like movies and books as well. The criterium to judge should be the same for them.
But I feel you are underestimating how much people care about the artist as well. Could AI get that level of support from people knowing its not a real human being? I don't think so. I mean most artist stories are also fabricated but if you explictly tell people hey this is like not real at all and fully ai but its a rapper from this particular city who grew up struggling with this particular issue so people from that city and facing those issues could relate to him probably, dont think it will work because people will still know its a bot pretending to be from their city and facing their particular issue.
Lastly i also generally public likes authenticity and the idea of an algorithm just spamming music is not very inspiring in that regard. Maybe in a blind test, they wouldn't even be able to tell which track is ai and which is not but the moment you tell them which one it is, they will likely think its bad, because who made the art completely changes the prespective of the consumer
> Spotify's personalization exemplifies invisible AI done right.
Ironically, I can't think of a platform that does personalization worse. Not only does it regularly surface music I don't like, it surfaces the exact same music I don't like over and over again. I don't know whether this is bad recommendation or their pushing music on me (ie payola) but it's supremely irritating.
I would have thought that adding semantic search to a photos app might be a better example of good AI, not these bolted-on-top examples.
Ironic that this very article has been partially written with AI... kind of lost the drive to read it
Yep. And if you want to integrate AI into your products, like we are currently doing, you are best off just delivering features / products and concealing the fact it's "AI" instead of putting it front and centre. AI has a rather tarnished brand amongst many groups of people.
How much of that is just pattern recognition?
I dislike AI because the result has consistently been bad. The most enthusiastic AI co-workers are producing garbage at a 100x rate, while the non-enthusiastic responsible ones are left reading and reviewing it.
More subtle thought-out AI tends to work better than AI for the sake of it anyway. If you deliver something useful, people will end up using it. A lot of current AI use is not particularly useful though.
I'd say "entirely", not "partially". Flagged the article.
Remember to flag it, I've been doing this consistently with AI slopicles. Eventually we can build a social consensus around this and stop having to see them.
It's kind of ironic, if this article was written about programming, it would be immediately flagged as being too thin and written almost completely with AI.
It's not just AI first companies..there's a place near me that tried to usr AI images in their social media ads. Completely blew up their relationship with the community. People would have preferred a cell phonr picture of their actual food to an AI generated approximation..
I've noticed this happening more and more! Mainly with restaurants! There's an indian place just down the street from my place with a menu composed of entirely (badly) AI generated curry pictures in the window. It turns everyone off eating there because it feels like they're somehow lying about what they sell. I get that curry is probably hard to photograph unprofessionally and still make it look appetizing, but I would've preferred
over"Consumers have developed pattern recognition for AI-generated content."
The irony of reading an article that talks about AI slop that clearly seems to have been written by AI. Hey, I could be completely wrong, and it wasn't, but there are so many flags.
Do I care? Not really, but whoever wrote this is right. I guess we developed a pattern recognition for these things
Agreed. But it's also indicating the current crop of AI generates certain patterns that are detectable.
What if it advances to a point where people can't? Chances are that's anywhere between "we're already past that point" and "could be a decade or more away" depending on the type of work.
Imho AI is ultimately just a tool. Use where it improves things, avoid where it doesn't. Right now companies seem to try & shove it everywhere. Which clearly isn't working.
It sure is a silver bullet ... to end your business.
It almost feels like AI is a hype and is used by the NFT/blockchain/crypto/web3 grifters as yet another vehicle to grift people right into their graves. Fun times.
AI as a term has always been a marketing tool aimed more at investors and executives than users.
Always? Maybe for the past 15-20 years but for the 50 years before that there wasn't much investment or interest from executives. AI was mostly about research and niche applications.
AI has always been a marketing term. As soon as it has a useful application it is named around its utility. Search, recognition, generation, classification, etc. AI is the term you use when you can't easily describe what you're doing.
It's true that in the early years, much of the investment was public money—but the above still stands.
This is true and is the reason I personally have always hated the term. This time around things are different. This time we also hate the technology.
Personally I used to love Markov Chains, Kalman filters, Supervised learning, etc. and resent the fact that people called it AI. But LLM has nothing but hate from my heart. Ironically, I love that people call it AI, and do so enthusiastically my self. A bad and hyped up technology deserves a bad and hyped up name.
the big far-right party in Austria recently did an AI poster for their 70 year celebration, and there really were 2 type of responses
- not their target group, thinking it was cringe and boomer-ish
- their uncritical target group, who loved a polished picture of blonde people
"When consumers believe emotional marketing communications are written by AI rather than humans, they judge them as less authentic, feel moral disgust and show weaker engagement and purchase intentions. This happens even when the content is otherwise identical."
Well, in general I do not care either way. I regard all ads as propaganda that attempts to steal my time. However had, even then it is indeed true that AI just is an additional annoyance factor, because it means that no real human really invested time - just AI slop that is spammed down onto people, and wastes their time. So I don't agree with the premise in the article to begin with, but most assuredly it is also true that AI slop just is pissing off people. I am noticing this on youtube too and although I don't have data, it seems that enough people were annoyed that the no-AI movement gained more grounds in the last some weeks. Hopfully we'll eventually reach AI extinction - not likely to happen, since some humans are already addicted to AI (see all "contributed with claude" on github spam), but I regard this as a noble goal. Rid this world of AI.
[flagged]
> Soon - and even now - there are people getting away with using it that you guys can’t detect.
Sure, but that seems largely irrelevant to what's being discussed here.
The public in general is anti-AI, so any marketeer thinking that associating AI with their brand is a positive is misreading the room, and any thinking that the public are unable to detect it most of the time, especially in something like a video, are just wrong.
Is that going to change - will AI-generated content become harder to detect? Sure, but "we know the public hate this, but we can con them" doesn't really sound like a recipe for success. In a world increasingly awash with AI generated content, it's easy to predict that human-generated content and interaction is going to be increasingly valued, and smart companies will realize this.
[flagged]
Lot of nobodies now think they are equal to ppl with skill and talent because intelligence has been democratized ( according to them) .
Most ai hypers fall into this category. They see any critism of ai as support of their current position at bottom of the talent stack.
show us what you built from this amazing technology, broski. you are still at the bottom. Bitching, whining and fantasy isnt your ticket to the top
> Do HN luddites realize the technology is here to stay?
> Soon - and even now - there are people getting away with using it that you guys can’t detect.
It's very strange to me that this is the pro-AI argument. It's not that things will be better, its that they can't be fought effectively.
The inventors of the printing press and the steam engine and the internet--whatever parallel you want--had a vision of improvement, making things better. AI proponents seem to want to make things worse, gleefully, but not be criticised for it.
Even if you're completely correct, that's not a good starting point.
They're making things worse, but think about how much more of it they are making!
Strong disagreement on this one, I think the coverage has proven that the current crop if anti-LLM sentiment is a much bigger thing then whatever is happening on this website.
In the case of AI generated imagery, maybe it will get to the point where people can't tell just by looking at it. If it does, at what cost? It seems to me that token cost is rising, I can't help but wonder if hiring an illustrator might end up being less expensive in the long run.
[dead]
Consumers and gamers are already alienated. If the price of RAM keeps going up I'm not sure much technology will actually be here to stay. What's the point of AI "art" anyway if everyone needs $10,000 devices to interact with the content? AI isn't going to raise wages, it's just going to enable people to produce more for the same wage.
[dead]
Do HN luddites realize the technology is here to stay?
Oh, I get that. LLMs are here to stay, like pyramid and Ponzi schemes, like Bitcoin and NFTs. It will be here until it isn't or we aren't. (And it just may help with that last.)
> Soon - and even now - there are people getting away with using it that you guys can’t detect.
hope this isnt some smartass selfrefretial joke
[dead]
You guys want something real to bitch about?
Look no further than the censorship in this thread! Why did mods delete half the replies and then ban the account?
You heard it from Tupac - this site has bigger problems than AI
LLMs are here to stay and I'm using them daily out of necessity and out of convenience too. I'm paying money for it too. It is a useful tech. Still, I hate it in general and the sociopaths promoting it in particular, and I foresee a great deal of issues for humanity because of that. Not catastrophic, just boringly bad. Like many humans going broke and jobless (booooring), or humans losing entire activities they actually liked to do to robots (booooring). It's sarcasm, if you didn't get it btw. I feel like I need to spell out a lot of things explicitly nowadays.
Like for example a lot of legitimate and useful companies are completely dependent on the advertisement, which is a net negative absolutely vile and disgusting invention. But we all collectively picked option to defect, and not because we are bad humans, and are paying for that decision collectively too.
> Including young illustrators on Twitter with a whole audience convinced they are drawing by hand.
that's called being a grifter not an illustrator