I don't really agree with this. The premise is that we should judge books (etc) by their covers because there's just too much information to do better. They're conflating AI capabilities with the...
I don't really agree with this. The premise is that we should judge books (etc) by their covers because there's just too much information to do better.
Many suggest AI as a solution to this filtering problem. Let the machine read and watch everything, recommending only what's truly valuable. Sounds promising, but in practice, it's deeply flawed.
Look at YouTube, for example. The algorithm doesn't prioritize what's important or valuable to you. It optimizes for engagement, which often translates to outrage, sensationalism, or endless distraction. AI serves the platform’s goals, not yours.
They're conflating AI capabilities with the goals of the people running it, who are people who have mastered the game of capitalism. That's not an AI limitation, just a problem of how we've built a system that rewards (and therefore necessitate) optimizing every technology for extracting wealth, and does not reward ethics or cautious experimentation.
And they may also be conflating LLMs with human built algorithms.
I agree with (and have written much about) how deeply flawed The Algorithm is on all major platforms, and how it damages is to individuals and society. But that's a very specific application of technology that can, to be honest, fairly easily done away with.
I'm skeptical that AI will ever be built in a way that prioritizes human needs over corporate needs. But I feel compelled to bring up the possibility every time this comes up.
I agree with you, and I've made similar comments on designing algorithms to benefit users. The trouble comes in primarily when the site owners exclusively control the levers that run the...
I agree with you, and I've made similar comments on designing algorithms to benefit users. The trouble comes in primarily when the site owners exclusively control the levers that run the algorithm.
Somewhat optimistically, I'd like to wrest that control away from site owners and back into the hands of users. Let us decide what we want to see in our recommendation engines, timeline sorting, and video highlights. These aren't inherently evil features, but they're a lot more useful to us when they're not prioritizing engagement, time on page, and active user counts.
Does doing this solve outrage, sensationalism, and other garbage? Not at all. People eat that stuff up, and if we blame their existence entirely on evil corps then we're being a bit naive. But many of us are drowning online right now - in information, and misinformation - and letting us slow the taps would give some of us the chance to breathe.
Fully agree. Ultimately these algorithms should be transparent. They represent countless decisions being made on our behalf by a business that has no interest in our well-being. Even if they're...
Fully agree. Ultimately these algorithms should be transparent. They represent countless decisions being made on our behalf by a business that has no interest in our well-being. Even if they're truly doing their best to do what is good for us, without transparency, it really can't be trusted.
I'm not sure what transparency looks like here, but I know it's solvable .
Not sure if I'm just playing devil's advocate here, or just being contrarian to you, but I'm fairly confident that what you've labelled as the premise is meant to be a tongue-in-cheek headline....
Not sure if I'm just playing devil's advocate here, or just being contrarian to you, but I'm fairly confident that what you've labelled as the premise is meant to be a tongue-in-cheek headline. The points in the article that are raised: us filtering the huge number of options are valid, for better or worse. We do it every single day. We may only be looking through the comedy tv shows category on Netflix for something that strikes our fancy. I have personally given preference to a resume with firstnamelastname@gmail over hotdoglover420@hotmail. Furthermore, I have chosen an article with a silly yet somewhat clickbait title that takes less than 10 minutes to read over the article about AI turning a man into a superhero because that one is estimated to take over 30 minutes. I can't get 30 minutes without interruption, and if I could I would read a book instead, because my time is so limited.
I judge books by their covers all the time. If I’m browsing the sci-fi section in the bookshop I’ll look for the big space ships and read the blurb on the covers to make choices. Not sure what...
I judge books by their covers all the time. If I’m browsing the sci-fi section in the bookshop I’ll look for the big space ships and read the blurb on the covers to make choices.
I'm the same way when I'm browsing for books if I'm not familiar with the author already. There is also a reason people spend money on cover art, to catch someone's eye who may be interested in...
I'm the same way when I'm browsing for books if I'm not familiar with the author already. There is also a reason people spend money on cover art, to catch someone's eye who may be interested in that particular style of covers.
I also know that I'm not as big of a fan of some cover trends, which I think I mentioned in a previous thread. Taking a look at The Wheel of Time, this Reddit thread highlights different covers for the books. I'm a big fan of the original covers for the books, and they were one of the reasons when my friend showed them to me all those years ago that I put them on my to read list. If I had no context for the series and saw the more minimalist covers at the books store they would not have caught my eye.
I think the change in cover style for Wheel of Time can be attributed to a changing purpose for the books. Everyone who is going to buy them is going to be buying them already knowing what they're...
I think the change in cover style for Wheel of Time can be attributed to a changing purpose for the books. Everyone who is going to buy them is going to be buying them already knowing what they're getting into, and so it's more important for them to look good on the shelf than to draw in new buyers. New books still have flashy covers.
I feel like with a series as large as Wheel of Time, that can make sense. Tastes differ, of course, and the newer covers aren't something I'd want to put on my bookshelf. For me, the same desire...
I feel like with a series as large as Wheel of Time, that can make sense. Tastes differ, of course, and the newer covers aren't something I'd want to put on my bookshelf. For me, the same desire of wanting them to attract attention on the shelf or be as enticing as I remember them being lets me introduce them to others who may not have read the series.
This is completely a taste thing, some people like their books with a more minimalist design and if they're reading I'm glad for it! I'm not a purest on one particular cover art for a series. I enjoy seeing different regional covers or different editions, and if I had the money for it, I would probably buy multiple copies of some of my books for the cool covers! In the vain of Wheel of time, something like this collectors edition stands out to me as looking fantastic as a set, even if the individual book covers aren't as distinctive.
There are a few books I've told my wife I'd like to get a fancy edition of one day (mainly Tolkien's works, especially since the various 100 year anniversary editions are going to pop up during my lifetime).
I think it's an okay way to approach media, as a baseline, with the realization there will be exceptions to the rule sometimes. I find it to be a decent guide when it comes to music. I avoid music...
I think it's an okay way to approach media, as a baseline, with the realization there will be exceptions to the rule sometimes.
I find it to be a decent guide when it comes to music. I avoid music with AI visuals by default. Also if the art, band logo, or aesthetics in general seem low quality often the music follows and most of the time I don't bother digging deeper to discover the exception.
The exceptions exist, though I find visuals to be "part of the package" when it comes to music, so even if the music is good but other parts are bad, in general it makes the overall expression feel lower quality to me, and less chance of getting my support. Especially because there's often very little chance I'm going to promote (through web, merch, linking to friends, etc) a band that sounds good but looks awful.
Every time I round-up my favorite weekly releases of each week of the year, I, not exaggerating, end up looking at hundreds/thousands of band names, cover art, album titles, etc and judging by the cover is my "first line of defense" filter for finding good things, and it generally steers me well
I recently judged a book by its cover. The Catalysts by Amanda Wick has a book cover that looks like a direct rip off of the Rush album cover for Vapor Trails. And the book cover supposedly had an...
I recently judged a book by its cover. The Catalysts by Amanda Wick has a book cover that looks like a direct rip off of the Rush album cover for Vapor Trails. And the book cover supposedly had an attributed cover artist, not a publisher generic cover.
I don't really agree with this. The premise is that we should judge books (etc) by their covers because there's just too much information to do better.
They're conflating AI capabilities with the goals of the people running it, who are people who have mastered the game of capitalism. That's not an AI limitation, just a problem of how we've built a system that rewards (and therefore necessitate) optimizing every technology for extracting wealth, and does not reward ethics or cautious experimentation.
And they may also be conflating LLMs with human built algorithms.
I agree with (and have written much about) how deeply flawed The Algorithm is on all major platforms, and how it damages is to individuals and society. But that's a very specific application of technology that can, to be honest, fairly easily done away with.
I'm skeptical that AI will ever be built in a way that prioritizes human needs over corporate needs. But I feel compelled to bring up the possibility every time this comes up.
I agree with you, and I've made similar comments on designing algorithms to benefit users. The trouble comes in primarily when the site owners exclusively control the levers that run the algorithm.
Somewhat optimistically, I'd like to wrest that control away from site owners and back into the hands of users. Let us decide what we want to see in our recommendation engines, timeline sorting, and video highlights. These aren't inherently evil features, but they're a lot more useful to us when they're not prioritizing engagement, time on page, and active user counts.
Does doing this solve outrage, sensationalism, and other garbage? Not at all. People eat that stuff up, and if we blame their existence entirely on evil corps then we're being a bit naive. But many of us are drowning online right now - in information, and misinformation - and letting us slow the taps would give some of us the chance to breathe.
Fully agree. Ultimately these algorithms should be transparent. They represent countless decisions being made on our behalf by a business that has no interest in our well-being. Even if they're truly doing their best to do what is good for us, without transparency, it really can't be trusted.
I'm not sure what transparency looks like here, but I know it's solvable .
Not sure if I'm just playing devil's advocate here, or just being contrarian to you, but I'm fairly confident that what you've labelled as the premise is meant to be a tongue-in-cheek headline. The points in the article that are raised: us filtering the huge number of options are valid, for better or worse. We do it every single day. We may only be looking through the comedy tv shows category on Netflix for something that strikes our fancy. I have personally given preference to a resume with firstnamelastname@gmail over hotdoglover420@hotmail. Furthermore, I have chosen an article with a silly yet somewhat clickbait title that takes less than 10 minutes to read over the article about AI turning a man into a superhero because that one is estimated to take over 30 minutes. I can't get 30 minutes without interruption, and if I could I would read a book instead, because my time is so limited.
I judge books by their covers all the time. If I’m browsing the sci-fi section in the bookshop I’ll look for the big space ships and read the blurb on the covers to make choices.
Not sure what else could be done really.
I'm the same way when I'm browsing for books if I'm not familiar with the author already. There is also a reason people spend money on cover art, to catch someone's eye who may be interested in that particular style of covers.
I also know that I'm not as big of a fan of some cover trends, which I think I mentioned in a previous thread. Taking a look at The Wheel of Time, this Reddit thread highlights different covers for the books. I'm a big fan of the original covers for the books, and they were one of the reasons when my friend showed them to me all those years ago that I put them on my to read list. If I had no context for the series and saw the more minimalist covers at the books store they would not have caught my eye.
I think the change in cover style for Wheel of Time can be attributed to a changing purpose for the books. Everyone who is going to buy them is going to be buying them already knowing what they're getting into, and so it's more important for them to look good on the shelf than to draw in new buyers. New books still have flashy covers.
I feel like with a series as large as Wheel of Time, that can make sense. Tastes differ, of course, and the newer covers aren't something I'd want to put on my bookshelf. For me, the same desire of wanting them to attract attention on the shelf or be as enticing as I remember them being lets me introduce them to others who may not have read the series.
This is completely a taste thing, some people like their books with a more minimalist design and if they're reading I'm glad for it! I'm not a purest on one particular cover art for a series. I enjoy seeing different regional covers or different editions, and if I had the money for it, I would probably buy multiple copies of some of my books for the cool covers! In the vain of Wheel of time, something like this collectors edition stands out to me as looking fantastic as a set, even if the individual book covers aren't as distinctive.
There are a few books I've told my wife I'd like to get a fancy edition of one day (mainly Tolkien's works, especially since the various 100 year anniversary editions are going to pop up during my lifetime).
I think it's an okay way to approach media, as a baseline, with the realization there will be exceptions to the rule sometimes.
I find it to be a decent guide when it comes to music. I avoid music with AI visuals by default. Also if the art, band logo, or aesthetics in general seem low quality often the music follows and most of the time I don't bother digging deeper to discover the exception.
The exceptions exist, though I find visuals to be "part of the package" when it comes to music, so even if the music is good but other parts are bad, in general it makes the overall expression feel lower quality to me, and less chance of getting my support. Especially because there's often very little chance I'm going to promote (through web, merch, linking to friends, etc) a band that sounds good but looks awful.
Every time I round-up my favorite weekly releases of each week of the year, I, not exaggerating, end up looking at hundreds/thousands of band names, cover art, album titles, etc and judging by the cover is my "first line of defense" filter for finding good things, and it generally steers me well
I recently judged a book by its cover. The Catalysts by Amanda Wick has a book cover that looks like a direct rip off of the Rush album cover for Vapor Trails. And the book cover supposedly had an attributed cover artist, not a publisher generic cover.