I read the article and it conveniently glosses over the fact that this will be a problem in the fairly near future for those who are using AI because the use of AI keeps a person on the page longer but doesn’t generally encourage them to view ads or actually click through to products or services. Those business are ad aggregation companies by default for the most part, and they aren’t gonna to survive without clickthroughs, not can the survive off ad revenue that doesn’t exist because people aren’t looking at the ads or clicking through to the products. The AI summary is the same problem that they already gave themselves once when they were trying to make search results more efficient in like 2018. You’d Google, and they would answer the query without you having to click a link at all. You’d get your answer and close the page.
Also, once they’ve starved themselves of user created data (as is posited by the article), they won’t have anything to keep feeding these models.
Those business are ad aggregation companies by default for the most part, and they aren’t gonna to survive without clickthroughs
Yes, this is it exactly. The web pages that depend on ad revenue are the ones in trouble here. They’re being undercut by pages that give people the information they want without going through all that stuff.
You’re confidently predicting that the AI summarizers are going to fail somehow, and then everything will just magically go back to the way they were. I suppose that’s a reassuring thing to believe. Why should I believe it, though? The AI genie is out of the bottle. I can run one locally on my computer if I want. All the existing online summarizers could go bankrupt tomorrow and I’ll still be able to get an AI to distill the information I want from the morass of ads and engagement-harvesting click farms.
This is me quoting my own comments on this article from another post on Lemmy:
I don’t like the way this article is written. There are concepts that it tries to convey that have major caveats it glosses over. Additionally it posits some ideas for alternatives that aren’t new currencies and doesn’t explain how most of them would work. It also seems to ignore the fact that content creators very often get paid in ad revenue by the very same companies that are exacerbating this problem with their GenAI models, as well as companies that are being hit hard by the lack of actual ad generated revenue due to loss of clickthroughs and impressions.
That being said it does actually somewhat explain a lot of the problem with the internet being sustained via ad revenue and ads.
Several of the companies who’s business model is built around ad aggregation are either investing in or developing/have launched GenAI products that are in opposition with their current business model.
They seem content at the moment to starve other places on the internet of the very ad revenue they rely on to make money. This will hurt them in the long run but they are focused on the short term profits they will make in the meantime and they do not seem concerned about the future so long as they can be seen to be on the cutting edge of the new technology.
I don’t really know if this will lead to a downturn in creator made content. A lot of paid creators are so invested in that eco system that they’d rather hop from one service to the next forever than give it up and go get a 9-5.
The pay as you crawl system is going to be difficult to implement, especially when crawlers already ignore the .txt file. The startups are not in a position to necessarily pay to license data and I question if they’d be able to pay as they crawl either. Meaning there will be big conglomerate gate keepers like Meta and Google and MS. The pay as you crawl system also only works if it’s regulated in some way so that normal users and small creators don’t get caught up in being victimized by bots/crawlers ignoring such rules or laws, with those victims unable to have their case taken seriously or heard at all.
As for determining where the information came from and providing attribution. Most people still aren’t going to click through to those pages. This is in part because a lot of them don’t want to see ads in the first place (for security reasons and because ads are an imposition on their increasingly limited time, energy, and attention). It’s also because they already have the information they need. You don’t care if Wikipedia gets your ad revenue so long as you can prove you were right about Brad Pitt’s height or his first job to your friend you made that bet with at the bar last night.
They say sources would be compensated. By who? And how? We have already established that people don’t think there’s a lot of value in paying for chatbots. The vast majority of Gen AI LLM users have shown (through polling, and introductory costs that go up in price later) that they aren’t interested in and don’t find value in pay for them. So conglomerates (many of whom run chatbots at a loss) would be on the hook both for paying for their crawlers and for providing such services to their consumers (corporate or not)? That most definitely is not sustainable.
The other option is licensing but a lot of data has already been crawled and continues to be crawled without licensing or compensation.
I’m not sure that changing this business model will lead to anything good.
Edit: There’s also the problem with DDOS attacks for smaller websites that get crawled. Lemmy has seen this first hand. It’s not the intention of the crawlers to overload the servers but there’s so many of them and the number keeps growing. There’s a whole lot of other issues besides the ones in this comment too.
I read the article and it conveniently glosses over the fact that this will be a problem in the fairly near future for those who are using AI because the use of AI keeps a person on the page longer but doesn’t generally encourage them to view ads or actually click through to products or services. Those business are ad aggregation companies by default for the most part, and they aren’t gonna to survive without clickthroughs, not can the survive off ad revenue that doesn’t exist because people aren’t looking at the ads or clicking through to the products. The AI summary is the same problem that they already gave themselves once when they were trying to make search results more efficient in like 2018. You’d Google, and they would answer the query without you having to click a link at all. You’d get your answer and close the page.
Also, once they’ve starved themselves of user created data (as is posited by the article), they won’t have anything to keep feeding these models.
Yes, this is it exactly. The web pages that depend on ad revenue are the ones in trouble here. They’re being undercut by pages that give people the information they want without going through all that stuff.
You’re confidently predicting that the AI summarizers are going to fail somehow, and then everything will just magically go back to the way they were. I suppose that’s a reassuring thing to believe. Why should I believe it, though? The AI genie is out of the bottle. I can run one locally on my computer if I want. All the existing online summarizers could go bankrupt tomorrow and I’ll still be able to get an AI to distill the information I want from the morass of ads and engagement-harvesting click farms.
This is me quoting my own comments on this article from another post on Lemmy:
I don’t like the way this article is written. There are concepts that it tries to convey that have major caveats it glosses over. Additionally it posits some ideas for alternatives that aren’t new currencies and doesn’t explain how most of them would work. It also seems to ignore the fact that content creators very often get paid in ad revenue by the very same companies that are exacerbating this problem with their GenAI models, as well as companies that are being hit hard by the lack of actual ad generated revenue due to loss of clickthroughs and impressions.
That being said it does actually somewhat explain a lot of the problem with the internet being sustained via ad revenue and ads.
Several of the companies who’s business model is built around ad aggregation are either investing in or developing/have launched GenAI products that are in opposition with their current business model.
They seem content at the moment to starve other places on the internet of the very ad revenue they rely on to make money. This will hurt them in the long run but they are focused on the short term profits they will make in the meantime and they do not seem concerned about the future so long as they can be seen to be on the cutting edge of the new technology.
I don’t really know if this will lead to a downturn in creator made content. A lot of paid creators are so invested in that eco system that they’d rather hop from one service to the next forever than give it up and go get a 9-5.
The pay as you crawl system is going to be difficult to implement, especially when crawlers already ignore the .txt file. The startups are not in a position to necessarily pay to license data and I question if they’d be able to pay as they crawl either. Meaning there will be big conglomerate gate keepers like Meta and Google and MS. The pay as you crawl system also only works if it’s regulated in some way so that normal users and small creators don’t get caught up in being victimized by bots/crawlers ignoring such rules or laws, with those victims unable to have their case taken seriously or heard at all.
As for determining where the information came from and providing attribution. Most people still aren’t going to click through to those pages. This is in part because a lot of them don’t want to see ads in the first place (for security reasons and because ads are an imposition on their increasingly limited time, energy, and attention). It’s also because they already have the information they need. You don’t care if Wikipedia gets your ad revenue so long as you can prove you were right about Brad Pitt’s height or his first job to your friend you made that bet with at the bar last night.
They say sources would be compensated. By who? And how? We have already established that people don’t think there’s a lot of value in paying for chatbots. The vast majority of Gen AI LLM users have shown (through polling, and introductory costs that go up in price later) that they aren’t interested in and don’t find value in pay for them. So conglomerates (many of whom run chatbots at a loss) would be on the hook both for paying for their crawlers and for providing such services to their consumers (corporate or not)? That most definitely is not sustainable.
The other option is licensing but a lot of data has already been crawled and continues to be crawled without licensing or compensation.
I’m not sure that changing this business model will lead to anything good.
Edit: There’s also the problem with DDOS attacks for smaller websites that get crawled. Lemmy has seen this first hand. It’s not the intention of the crawlers to overload the servers but there’s so many of them and the number keeps growing. There’s a whole lot of other issues besides the ones in this comment too.