Posting internet comments is the one form of media that proletarians can easily participate in to reach the masses without outright gatekeeping by an editor. Yeah yeah, mods & admins can remove comments/posts, and algorithms can fuck with their reach, but generally, discussion can exist without too much intervention.
It was absolutely apparent that disrupting this was an explicit goal of LLMs from their initial outset.
This creates an asymmetrical playing field where corporations and NED orgs can spend lots of money on GPU farms to push their messages, but prole groups cannot nearly as easily.
LLMs also trivially allow these same efforts to interface with the only platforms uncolonized by corporate algorithm fuckery. So even if we built ourselves a life raft with the fediverse / Lemmy, we are still trivially subject to large scale influence operations, even if less valuable as a target. It seems like big tech is aware of efforts to avoid corporate control over public opinion, and LLMs are the immune system response to exert control on platforms they don’t own.
Posting internet comments is the one form of media that proletarians can easily participate in to reach the masses without outright gatekeeping by an editor
Be for real lol. You’re posting on one of the only places on the internet that isn’t completely owned by the bourgeoisie. Your reach is already limited by segregating yourself on this site. Pretending that posting on “the internet” (e.g. the biggest sites) which is mostly controlled by gigantic bourgeoisie interests and petite bourgeoisie interests is a purely proletarian act is silly. In practice “freedom to post” is ancillary to feeding you advertising. You are playing in their sandbox and have always been. The enshittification of the 2010’s blurred the lines between content and advertising intentionally because the smallest advertisers who refused to pay for ads already started doing that in the late 2000’s.
This creates an asymmetrical playing field where corporations and NED orgs can spend lots of money on GPU farms to push their messages, but prole groups cannot nearly as easily.
You’re just competing over different resources. Prior to the mechanization of this through any software you were competing for bodies doing the posting (whether the medium was internet, newspapers, leaflets, pamphlets, word of mouth etc doesn’t matter). With augmented mechanization prior to AI you were competing for bodies and compute. Now you can argue you mainly compete over compute. However because of the fun logic of capitalism, in the current state of capitalist organization proletarians can afford compute more than they can afford to feed other proletarians amortized over time. So this really only boils down to “humans are becoming obsolete in the posting wars”, which was already true.
The inherent contradictions in capitalism make it effectively impossible for this technology to be really gate kept from proles over time without a literal crackdown on ownership or development. The cat is so out of the bag there’d be a civil war if the bourgeoisie started drawing lines on who can own what things for the “greater good”. The US could not even keep a lid on encryption software over time, despite export controls for encryption still being a thing.
So even if we built ourselves a life raft with the fediverse / Lemmy, we are still trivially subject to large scale influence operations, even if less valuable as a target. It seems like big tech is aware of efforts to avoid corporate control over public opinion
This was already true before 2020. The type of software that major botnets/intelligence services/ad companies specializing in guerilla advertising is called a Persona Management System. These systems mechanized the dissemination of posting and conversation shaping across the internet. It has only really been in the last 8 years that Persona Management has been a thing that exists outside of intelligence adjacent circles.
What AI does is it makes running / developing Persona Management Systems easier. You no longer have to make a cohesive overarching strategy to make things feel organic for humans, you can simply be “good enough” with AI and that’s always what these systems campaigns aim for.
Don’t get me wrong this is an acceleration, but it’s not an acceleration that brings us to a “brave new world”. Yes the owners have new toys, but they haven’t been able to prevent us from using them. Likewise if this was a real “information war”, this becomes a war of logistics like every other war. It becomes necessary for proles to find a tactic that allows them to make the equivalent of a Hamas piss rocket or a Quds drone, something that is incredibly cheap but because of the capitalist state of the world requires a wasteful amount money to defend against. For every dollar in prole spend we make them spend $1,000 against it.
Smaller sites are already protecting themselves using proof of work systems like anubis to prevent AI crawling. Anubis is a great example of making them pay more. If Hexbear/Lemmy implemented anubis, it would be a slight annoyance for us an extra 1-2s browsing. However for an AI crawler running at scale to “take over the community” it would create a significant financial disadvantage because at scale the costs for running a hashcash for every interaction would balloon.
This isn’t even new for reddit. They seeded the original site back during the 2005-2007 years by paying people to run sockpuppets to simulate the network effect. They’d post about tech before, the subject matter they’re posting now is a level of degeneracy that only comes with an IPO.
Be for real lol. You’re posting on one of the only places on the internet that isn’t completely owned by the bourgeoisie. Your reach is already limited by segregating yourself on this site. Pretending that posting on “the internet” (e.g. the biggest sites) which is mostly controlled by gigantic bourgeoisie interests and petite bourgeoisie interests is a purely proletarian act is silly. In practice “freedom to post” is ancillary to feeding you advertising. You are playing in their sandbox and have always been.
I knew my comment would inevitably get this response.
Reaffirming that I do not believe that posting is praxis. With regard to bourgeois owned platforms, it is without question any serious revolutionary discussion that cannot be coopted will get squashed. Occasionally they fail to control the narrative, and the limited opportunities we have to sneak in counter narratives are more easily squashed with deployments of large scale LLM opposition. Obviously we should have no confidence of using their platforms to counter their interests to any effect besides finding like-minded people and taking interactions off their platforms.
What makes LLMs different from existing tools available to bourgeois platforms is the ability to extend into public spaces they do not directly control like the fediverse or BBS forums and image boards.
The inherent contradictions in capitalism make it effectively impossible for this technology to be really gate kept from proles over time without a literal crackdown on ownership or development.
Gatekeeping is not required, especially not preventing the purchase of GPUs.
But even taking account purchasing, corpos and NED orgs can benefit from economies of scale, whereas people going to Amazon or Best Buy will be limited to a handful of units at a time. This is the least significant limitation we face.
The asymmetry is not just the accumulation of GPUs, but also other resources like IP addresses and uniformity of deployment. Individual proles or small orgs will not be able to launch campaigns similar to how a NED funded org, tech PR team, or other multinational corporation. Our IPs and accounts will be blocked, theirs whitelisted. Our efforts will be disparate and unable to react to real-time information like sentiment analysis, theirs will. Any three letter org can probably go to Facebook or Google and be handed hundreds of thousands of real looking, aged accounts.
Per GPU, they will be able to spew more LLM bullshit than us given, even on our own platforms, because they have organizational advantages.
It becomes necessary for proles to find a tactic that allows them to make the equivalent of a Hamas piss rocket or a Quds drone, something that is incredibly cheap but because of the capitalist state of the world requires a wasteful amount money to defend against.
Strongly agree that something like this is needed. However, what enables the asymmetry here is a limitation on material resources and labor. No such limitation exists with digital systems that cannot be trivially copied and deployed by bourgeois interests.
Only idea I have for something like this is a system that requires user verification by existing members, possibly gatekept by irl meetings and capped at some rate per user. Few people use the fediverse because of tiny barriers to entry. No one would tolerate a system that imposes more burdens.
DDoS prevention solutions are probably capable of thwarting most technological solutions we can come up with to create our own asymmetries.
The type of software that major botnets/intelligence services/ad companies specializing in guerilla advertising was is called a Persona Management Systems.
Do you have any specific knowledge on the characteristics, capabilities, or implementation of these? I would be very interested in learning more.
Strongly agree that something like this is needed. However, what enables the asymmetry here is a limitation on material resources and labor. No such limitation exists with digital systems that cannot be trivially copied and deployed by bourgeois interests.
Only idea I have for something like this is a system that requires user verification by existing members, possibly gatekept by irl meetings and capped at some rate per user. Few people use the fediverse because of tiny barriers to entry. No one would tolerate a system that imposes more burdens.
DDoS prevention solutions are probably capable of thwarting most technological solutions we can come up with to create our own asymmetries.
The root of your perception here is correct, the bourgeoisie will always have “promoted posts” for their personal interests, but that is no different than any censor or editor that exists today.
You are also correct in the economies of scale, my point is that we have always been playing catch up and these tools allow us to catch up significantly as well. These tools raise our capability floor to heights that were not possible before.
Do you have any specific knowledge on the characteristics, capabilities, or implementation of these? I would be very interested in learning more.
Pega and Capado are the advertising ones. Sock puppet is a commercial offering targetted at OSINT, “to protect researcher identities”.
The knowledge we have of the US government ones are from things like Operation Earnest Voice where they contracted out development to Ntrepid.
Most software out there isn’t a highly capable consistent package. I’m sure the CIA or Palantir has one developed, but a lot of places just run developer tooling like Playwright to control Chrome. Prior to LLM’s any procedural text generation was done via Markov Chains and sentiment analysis thru keyword analysis (e.g. regex).
There are some OSS tools out there in various states of quality like:
I did a whole bunch of research on this as I thought it would be a fun challenge to build one, I wanted to make it OSS, but I decided against it after I generally thought about what “success” would look like, I was basically thinking of building a turnkey internet nuke.
Typically your keywords are “persona management”, “sock puppet tools” etc.
Thank you for this info. Lots of great launch points for looking deeper into this.
my point is that we have always been playing catch up and these tools allow us to catch up significantly as well. These tools raise our capability floor to heights that were not possible before.
Observation of the development of FOSS ecosystems seems to back this up and does give me some hope for the state of technology to build alternatives to capitalist run software.
We do have our own asymmetries like not having the profitability requirement and being able to onboard developers without the barriers the constraints of employment impose.
I did a whole bunch of research on this as I thought it would be a fun challenge to build one, I wanted to make it OSS, but I decided against it after I generally thought about what “success” would look like, I was basically thinking of building a turnkey internet nuke.
Curious about this if you can expand upon it without doxxing yourself.
Not sure if I am making the correct inference of what you are hinting at, but my initial thoughts are that something that creates a broad distrust of social media would result in a better world than a world where pre-existing comrades are the only people who have that distrust with no way of breaking through the astroturfed counter narratives.
Given FOSS bros tendency towards liberalism and reactionary ideology, open source would probably be a bad way to deploy something like this. If I had an implementation of “a turnkey internet nuke”, I’d be booking a flight to Beijing before releasing it on the open internet…or at least minimally build a cadre of capable comrades I could trust to use it responsibly.
Curious about this if you can expand upon it without doxxing yourself.
Often software success is measured by adoption. Let’s say tomorrow a turn key persona management system existed.
The most institutionalized levels of success possible is to be coopted by and contract out to an intelligence service (immediate pass, even the attention, immediate pass). However that means that such a system must unknowingly solve the problems of:
the features that are absolutely required for said service
the ease of migration of operations onto said service
be completely self evident to many at that service to be an answer for their problems
be a person that they could feel they could trust or manipulate into control
Very unlikely.
Another level of success is adoption in the general ecosystem. So now the DSA’s, PSL’s of the world can run persona management easily. Not really a big socialism success given you literally said it yourself it’s a tool that if proliferated creates a
“broad distrust of social media”.
So who are gonna be the biggest users of such a tool? Open your phone and look at your texts. That’s right the people who are shaking down grandparents, the people who are pretending to be a rich celebrity that has taken an interest in a lonely isolated person living in a rural community, scammers, flim flam artists, NFT enthusiasts, crypto bros.
The problem is that there is already a “broad distrust of social media”, that doesn’t stop humanity from using the communication tool, and that doesn’t lower the total incidences of exploitation that happen.
If I had an implementation of “a turnkey internet nuke”, I’d be booking a flight to Beijing before releasing it on the open internet
Lol China would not have you. This is a movie plot. They don’t want guys who make “rocking the boat technology” at all. You think they’d want some OSS enthusiasts running experiments with apes commenting on the net in their green zone? They have their own versions. They have their own trusted operators. This is like adoption with any other intelligence service.
In general the problem with this technology is literally like nukes. It shouldn’t exist. It’s easier to make social systems without its existence. This tech doesn’t change anything, it makes a better integrated version of desperate components of an existing system.
If tomorrow at the same time everyone woke up with a turnkey nuclear bomb, Israel would pre-emptively nuke Gaza.
I only hope that either capitalism crashes and makes this shit too expensive for them to do, or artists and writers succeed in their efforts on poisoning the algos
Our only hope is China maintaining course and leading in artificial intelligence.
IMO capitalism in the west cannot fall before the fall of bourgeois owned social media and big tech, so we are probably doomed forever.
But the future for the global south and AES is bright.
The only upside to the US trying to instigate a war in Taiwan is that if China or the US bombs TSMC, all of this becomes immediately financially infeasible.
This is the only way capitalism will fall in the imperial core, unless we wait decades for China to massively outpace our AI development, completely change their foreign policy, and deploy better LLMs in the west pushing revolutionary counter narratives.
Posting internet comments is the one form of media that proletarians can easily participate in to reach the masses without outright gatekeeping by an editor. Yeah yeah, mods & admins can remove comments/posts, and algorithms can fuck with their reach, but generally, discussion can exist without too much intervention.
It was absolutely apparent that disrupting this was an explicit goal of LLMs from their initial outset.
This creates an asymmetrical playing field where corporations and NED orgs can spend lots of money on GPU farms to push their messages, but prole groups cannot nearly as easily.
LLMs also trivially allow these same efforts to interface with the only platforms uncolonized by corporate algorithm fuckery. So even if we built ourselves a life raft with the fediverse / Lemmy, we are still trivially subject to large scale influence operations, even if less valuable as a target. It seems like big tech is aware of efforts to avoid corporate control over public opinion, and LLMs are the immune system response to exert control on platforms they don’t own.
Be for real lol. You’re posting on one of the only places on the internet that isn’t completely owned by the bourgeoisie. Your reach is already limited by segregating yourself on this site. Pretending that posting on “the internet” (e.g. the biggest sites) which is mostly controlled by gigantic bourgeoisie interests and petite bourgeoisie interests is a purely proletarian act is silly. In practice “freedom to post” is ancillary to feeding you advertising. You are playing in their sandbox and have always been. The enshittification of the 2010’s blurred the lines between content and advertising intentionally because the smallest advertisers who refused to pay for ads already started doing that in the late 2000’s.
You’re just competing over different resources. Prior to the mechanization of this through any software you were competing for bodies doing the posting (whether the medium was internet, newspapers, leaflets, pamphlets, word of mouth etc doesn’t matter). With augmented mechanization prior to AI you were competing for bodies and compute. Now you can argue you mainly compete over compute. However because of the fun logic of capitalism, in the current state of capitalist organization proletarians can afford compute more than they can afford to feed other proletarians amortized over time. So this really only boils down to “humans are becoming obsolete in the posting wars”, which was already true.
The inherent contradictions in capitalism make it effectively impossible for this technology to be really gate kept from proles over time without a literal crackdown on ownership or development. The cat is so out of the bag there’d be a civil war if the bourgeoisie started drawing lines on who can own what things for the “greater good”. The US could not even keep a lid on encryption software over time, despite export controls for encryption still being a thing.
This was already true before 2020. The type of software that major botnets/intelligence services/ad companies specializing in guerilla advertising is called a Persona Management System. These systems mechanized the dissemination of posting and conversation shaping across the internet. It has only really been in the last 8 years that Persona Management has been a thing that exists outside of intelligence adjacent circles.
What AI does is it makes running / developing Persona Management Systems easier. You no longer have to make a cohesive overarching strategy to make things feel organic for humans, you can simply be “good enough” with AI and that’s always what these systems campaigns aim for.
Don’t get me wrong this is an acceleration, but it’s not an acceleration that brings us to a “brave new world”. Yes the owners have new toys, but they haven’t been able to prevent us from using them. Likewise if this was a real “information war”, this becomes a war of logistics like every other war. It becomes necessary for proles to find a tactic that allows them to make the equivalent of a Hamas piss rocket or a Quds drone, something that is incredibly cheap but because of the capitalist state of the world requires a wasteful amount money to defend against. For every dollar in prole spend we make them spend $1,000 against it.
Smaller sites are already protecting themselves using proof of work systems like anubis to prevent AI crawling. Anubis is a great example of making them pay more. If Hexbear/Lemmy implemented anubis, it would be a slight annoyance for us an extra 1-2s browsing. However for an AI crawler running at scale to “take over the community” it would create a significant financial disadvantage because at scale the costs for running a hashcash for every interaction would balloon.
This isn’t even new for reddit. They seeded the original site back during the 2005-2007 years by paying people to run sockpuppets to simulate the network effect. They’d post about tech before, the subject matter they’re posting now is a level of degeneracy that only comes with an IPO.
Only halfway through this comment chain but it sucks that Persona Management Systems is a marketing term and not a standard JRPG combat mechanic
I knew my comment would inevitably get this response.
Reaffirming that I do not believe that posting is praxis. With regard to bourgeois owned platforms, it is without question any serious revolutionary discussion that cannot be coopted will get squashed. Occasionally they fail to control the narrative, and the limited opportunities we have to sneak in counter narratives are more easily squashed with deployments of large scale LLM opposition. Obviously we should have no confidence of using their platforms to counter their interests to any effect besides finding like-minded people and taking interactions off their platforms.
What makes LLMs different from existing tools available to bourgeois platforms is the ability to extend into public spaces they do not directly control like the fediverse or BBS forums and image boards.
Gatekeeping is not required, especially not preventing the purchase of GPUs.
But even taking account purchasing, corpos and NED orgs can benefit from economies of scale, whereas people going to Amazon or Best Buy will be limited to a handful of units at a time. This is the least significant limitation we face.
The asymmetry is not just the accumulation of GPUs, but also other resources like IP addresses and uniformity of deployment. Individual proles or small orgs will not be able to launch campaigns similar to how a NED funded org, tech PR team, or other multinational corporation. Our IPs and accounts will be blocked, theirs whitelisted. Our efforts will be disparate and unable to react to real-time information like sentiment analysis, theirs will. Any three letter org can probably go to Facebook or Google and be handed hundreds of thousands of real looking, aged accounts.
Per GPU, they will be able to spew more LLM bullshit than us given, even on our own platforms, because they have organizational advantages.
Strongly agree that something like this is needed. However, what enables the asymmetry here is a limitation on material resources and labor. No such limitation exists with digital systems that cannot be trivially copied and deployed by bourgeois interests.
Only idea I have for something like this is a system that requires user verification by existing members, possibly gatekept by irl meetings and capped at some rate per user. Few people use the fediverse because of tiny barriers to entry. No one would tolerate a system that imposes more burdens.
DDoS prevention solutions are probably capable of thwarting most technological solutions we can come up with to create our own asymmetries.
Do you have any specific knowledge on the characteristics, capabilities, or implementation of these? I would be very interested in learning more.
The root of your perception here is correct, the bourgeoisie will always have “promoted posts” for their personal interests, but that is no different than any censor or editor that exists today.
You are also correct in the economies of scale, my point is that we have always been playing catch up and these tools allow us to catch up significantly as well. These tools raise our capability floor to heights that were not possible before.
Pega and Capado are the advertising ones. Sock puppet is a commercial offering targetted at OSINT, “to protect researcher identities”.
The knowledge we have of the US government ones are from things like Operation Earnest Voice where they contracted out development to Ntrepid.
Most software out there isn’t a highly capable consistent package. I’m sure the CIA or Palantir has one developed, but a lot of places just run developer tooling like Playwright to control Chrome. Prior to LLM’s any procedural text generation was done via Markov Chains and sentiment analysis thru keyword analysis (e.g. regex).
There are some OSS tools out there in various states of quality like:
I did a whole bunch of research on this as I thought it would be a fun challenge to build one, I wanted to make it OSS, but I decided against it after I generally thought about what “success” would look like, I was basically thinking of building a turnkey internet nuke.
Typically your keywords are “persona management”, “sock puppet tools” etc.
Thank you for this info. Lots of great launch points for looking deeper into this.
Observation of the development of FOSS ecosystems seems to back this up and does give me some hope for the state of technology to build alternatives to capitalist run software.
We do have our own asymmetries like not having the profitability requirement and being able to onboard developers without the barriers the constraints of employment impose.
Curious about this if you can expand upon it without doxxing yourself.
Not sure if I am making the correct inference of what you are hinting at, but my initial thoughts are that something that creates a broad distrust of social media would result in a better world than a world where pre-existing comrades are the only people who have that distrust with no way of breaking through the astroturfed counter narratives.
Given FOSS bros tendency towards liberalism and reactionary ideology, open source would probably be a bad way to deploy something like this. If I had an implementation of “a turnkey internet nuke”, I’d be booking a flight to Beijing before releasing it on the open internet…or at least minimally build a cadre of capable comrades I could trust to use it responsibly.
Often software success is measured by adoption. Let’s say tomorrow a turn key persona management system existed.
The most institutionalized levels of success possible is to be coopted by and contract out to an intelligence service (immediate pass, even the attention, immediate pass). However that means that such a system must unknowingly solve the problems of:
Very unlikely.
Another level of success is adoption in the general ecosystem. So now the DSA’s, PSL’s of the world can run persona management easily. Not really a big socialism success given you literally said it yourself it’s a tool that if proliferated creates a “broad distrust of social media”.
So who are gonna be the biggest users of such a tool? Open your phone and look at your texts. That’s right the people who are shaking down grandparents, the people who are pretending to be a rich celebrity that has taken an interest in a lonely isolated person living in a rural community, scammers, flim flam artists, NFT enthusiasts, crypto bros.
The problem is that there is already a “broad distrust of social media”, that doesn’t stop humanity from using the communication tool, and that doesn’t lower the total incidences of exploitation that happen.
Lol China would not have you. This is a movie plot. They don’t want guys who make “rocking the boat technology” at all. You think they’d want some OSS enthusiasts running experiments with apes commenting on the net in their green zone? They have their own versions. They have their own trusted operators. This is like adoption with any other intelligence service.
In general the problem with this technology is literally like nukes. It shouldn’t exist. It’s easier to make social systems without its existence. This tech doesn’t change anything, it makes a better integrated version of desperate components of an existing system.
If tomorrow at the same time everyone woke up with a turnkey nuclear bomb, Israel would pre-emptively nuke Gaza.
I only hope that either capitalism crashes and makes this shit too expensive for them to do, or artists and writers succeed in their efforts on poisoning the algos
Our only hope is China maintaining course and leading in artificial intelligence.
IMO capitalism in the west cannot fall before the fall of bourgeois owned social media and big tech, so we are probably doomed forever.
But the future for the global south and AES is bright.
The only upside to the US trying to instigate a war in Taiwan is that if China or the US bombs TSMC, all of this becomes immediately financially infeasible.
This is the only way capitalism will fall in the imperial core, unless we wait decades for China to massively outpace our AI development, completely change their foreign policy, and deploy better LLMs in the west pushing revolutionary counter narratives.
The hope of getting a Chinese visa is the only thing keeping me going at this point.