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).
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.
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: https://github.com/dimensi0n/persona
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.