So far nothing is being done to manage the flood of #AISlop on the #Internet.
-
But everyone who shared receives a notification that you have edited the original post.
@uc While that is true, all the votes did not come back, only a smaller part, as far as I could tell (obviously not representative, but still).
-
@uc While that is true, all the votes did not come back, only a smaller part, as far as I could tell (obviously not representative, but still).
I don't think they're allowed a second vote, even if the votes are reset to zero. Not sure, but this is in the back of my mind.
-
@madeindex Most of the slop is motivated by money. Once you are not being paid by the click the motivation for the slop goes way down.
@gam3 Yeah that is true. The political etc. content seems to be a much smaller portion. So you are suggesting the ban of all ads on the web? Probably also unrealistic as it also finances a large portion of it, right?
-
@squareScarletCougar
The bane of my existence is that to promote the Fedi and other causes, I have to maintain accounts for this project on the dark side - as it's the best place to reach people who are not already agreeing hehe
@madeindex its no different from when we did propaganda against the Nazis https://youtu.be/nk1eWaic2-M?si=VdxUHRNgAcBIeIor
-
I don't think they're allowed a second vote, even if the votes are reset to zero. Not sure, but this is in the back of my mind.
@uc oh so you were not able to press vote again? interesting, thanks for telling me, maybe it's worth opening a Masto GitHub ticket for this?
-
So far nothing is being done to manage the flood of #AISlop on the #Internet.
The single way to "kind of" reduce it is by looking at the source (long established / people you know or that have a large following / are trusted).
Of course there is always change and that's a good thing, but this fundamentally changes things.
Will the internet as we know die because of #AI slop?
#LLM #artificialintelligence #artificial #digital #online #content #tech #technology #web #it #socialmedia #problem
@MadeInDex
if ai slop is taking over the internet then ai will choke on itsself. because ai feeding on the content of the internet and ruminating already processed ai slop will lead to a model collapse. so ai will either be totally easy to spot after this collapse, or the ai companies will have to develop a way to identify ai slop and avoid it. this ability will then become common knowledge and a kind of equilibrium will develop ...
#^https://www.nature.com/articles/s41586-024-07566-y -
@MadeInDex if ai slop is taking over the internet then ai will choke on itsself. because ai feeding on the content of the internet and ruminating already processed ai slop will lead to a model collapse. so ai will either be totally easy to spot after this collapse, or the ai companies will have to develop a way to identify ai slop and avoid it. this ability will then become common knowledge and a kind of equilibrium will develop ...
#^https://www.nature.com/articles/s41586-024-07566-y@jabgoe2089 Very interesting point, never even heard of this before, thanks for sharing! That would be extremely funny to watch
However they can always opt to exclusively train on old content up to like 2020 or something right?
-
@jabgoe2089 Very interesting point, never even heard of this before, thanks for sharing! That would be extremely funny to watch
However they can always opt to exclusively train on old content up to like 2020 or something right?
@MadeInDex sure, as long as they are using unprocessed content the ai development should be fine. but eg in 2030 we as consumers should be able to identify ai texts and images based on stuff before 2020
-
So far nothing is being done to manage the flood of #AISlop on the #Internet.
The single way to "kind of" reduce it is by looking at the source (long established / people you know or that have a large following / are trusted).
Of course there is always change and that's a good thing, but this fundamentally changes things.
Will the internet as we know die because of #AI slop?
#LLM #artificialintelligence #artificial #digital #online #content #tech #technology #web #it #socialmedia #problem
@madeindex "AI" already damaged the open internet because to produce slop you first need to pilfer every accessible pool of information, and that brings down servers
But there is no point crying, in a sense some of this disruption was inevitable, it's a tipping point created by all of us making a lot of good digital information, code, art etc. available.
It's been exploited by bad actors but we need to really think what good actors would do. Screaming "fire" is at best the first step.
-
@madeindex "AI" already damaged the open internet because to produce slop you first need to pilfer every accessible pool of information, and that brings down servers
But there is no point crying, in a sense some of this disruption was inevitable, it's a tipping point created by all of us making a lot of good digital information, code, art etc. available.
It's been exploited by bad actors but we need to really think what good actors would do. Screaming "fire" is at best the first step.
@openrisk Wow never even heard of that issue before, learning so many new things today, thank you very much for sharing that with me
My solution would be to ask or pay #Google to share their "data analyzation" algorithm:
https://mastodon.social/@madeindex/116204089351991213If anyone can still distinguish #AI content from "real" content, it's very likely Google, as that has been at the core of their business model for so long.
-
I don't think they're allowed a second vote, even if the votes are reset to zero. Not sure, but this is in the back of my mind.
@uc can you please confirm as a reply that you were unable to vote again, so I can open a ticket on the Masto GitHub?
-
@madeindex My experience too, so... Yes? Makes sense; it prevents users from manipulating poll results.
@jwcph were you unable to vote again in the poll after that? thanks
-
@uc can you please confirm as a reply that you were unable to vote again, so I can open a ticket on the Masto GitHub?
I just voted again, so perhaps that has been changed get.
-
I just voted again, so perhaps that has been changed get.
@uc ok, then I won't open one, might have just been an issue with your browser or something eh
-
@gam3 Yeah that is true. The political etc. content seems to be a much smaller portion. So you are suggesting the ban of all ads on the web? Probably also unrealistic as it also finances a large portion of it, right?
@madeindex YouTube is very expensive and runs on ad revenue, so I think it is the test case. The shorts, which are more click driven that time driven seem to be much worse than the long format videos.
Competition is likely the answer though. If YouTube gets to bad people will start paying for Nebula or figure out a way to pay for a PeerTube server or somesuch.
But it't the pay per click that is the problem, not the ad IMHO. -
@madeindex YouTube is very expensive and runs on ad revenue, so I think it is the test case. The shorts, which are more click driven that time driven seem to be much worse than the long format videos.
Competition is likely the answer though. If YouTube gets to bad people will start paying for Nebula or figure out a way to pay for a PeerTube server or somesuch.
But it't the pay per click that is the problem, not the ad IMHO.@gam3 but But why is AI PPC driven? Somehow seems to me like there are many other things it's used for, such a AD revenue, Affiliate Marketing, Promo, Propaganda and such?
-
@openrisk Wow never even heard of that issue before, learning so many new things today, thank you very much for sharing that with me
My solution would be to ask or pay #Google to share their "data analyzation" algorithm:
https://mastodon.social/@madeindex/116204089351991213If anyone can still distinguish #AI content from "real" content, it's very likely Google, as that has been at the core of their business model for so long.
I have been running a wiki for more than a decade for people to use as they see fit. Its no longer possible because for each person visiting you have ten thousand bots scrapping indiscriminately. Effectively I must pay for a server for the "AI" bros to build LLMs to serve the same information in chatbot format (and collect payment
)The wiki is now behind a login. But even that will not survive. LLM's impersonating real people (e.g by filling forms) is getting easier by the day.
-
I have been running a wiki for more than a decade for people to use as they see fit. Its no longer possible because for each person visiting you have ten thousand bots scrapping indiscriminately. Effectively I must pay for a server for the "AI" bros to build LLMs to serve the same information in chatbot format (and collect payment
)The wiki is now behind a login. But even that will not survive. LLM's impersonating real people (e.g by filling forms) is getting easier by the day.
@openrisk sorry to hear that, how much has the traffic load reduced ever since you implemented that feature?
