ai propaganda for you and me

fancy slopping you here

Slop has an intentional meaning which is more subtle than the common verbiage of the vernacular of the online disgruntled discourse cohorts:

Slop presupposes a user of AI or some auto-generation method is using the auto-generated content, be it text, audio, video, vegetable, or mineral as a substitute for their own effort, work, output, judgment, taste, and style.

but what if you aren’t using “ai slop” to substitute for your own ability, but to extend or enhance or resolve concepts you could contain but not express yourself (due to time or physical limitations) directly?

let’s articulate slop qua non-slop: I think there’s two categories of non-slop: broadcast vs functional:

  • broadcast non-slop: Something by definition cannot be AI slop if you do not intend for it to be consumed seriously as misrepresenting your own learned/trained capabilities in a medium of understanding experience or enjoyment or units of economic advantage.
  • functional non-slop: Something by definition cannot be AI slop if you “generated” it and it works with no fewer defects per size (per loc/file/plan/etc) than if you did the work “by hand.” i.e. if I make (and audit/edit/refine/architect/review/test/manage/iterate) a 300,000 line distributed system in 1 month using ai code robot generator helper bro pals and it only has 10 defects, where, if I tried to do it by hand, it would have taken me personally 7 years and had 1,000 problems instead, then it is still non-slop.

skip to the punchline: facetious propaganda video ranking app

The nuance of what it means to be consumed seriously is delicate because “seriously” can also have multiple dimensions: posting to platforms which themselves are natural slop magnets like toktics and ugrams and others, which, even if you are posting insincerely while knowing your insincere posts will garner some benefit for yourself (as far as like/view count revenue or advertising revenue; see: those slop YouTube children’s channels from bulk Eastern European or Russian content farm factories, which are just endlessly broken brainrot+ipad-child loop stuff).

Low quality posting floods were pre-AI slop slop, right? having auto-generated management processes to crank out as much content to be consumed with no artistic intent just for boosting ad revenue from attention fraud at scale. Also, things like the infinite life hacks or other channels which just create completely implausible dopamine hits with no intention of being taken for anything other than economic advantage for ad boosting click revenue over time. Those are all pure slop/brainrot no matter the generating medium itself (and speaking of “medium” it has always been full of slop like “datasceince” blogs doing seo spam garbage or wider “the death of google” infinite “geeksforgeeks” seo spam slop ad fraud sites etc – no ai was even needed on those, just infinite exploiting content farms from low income countries to pollute the entire world all at once).

So if you break down slop versus not slop:

  • slop: you are intentionally misrepresenting yourself or your abilities or the effort you put into something to pass off auto-generated work as something you are claiming to, but have not, conceived of, edited, produced, and submitted into the Global Human Record as your own work.
  • slop: Anything where you are deriving a personal advantage from generated output, which you yourself are not taking seriously, or do not care about the end result or impact of any usage or consumption of your creations at all. Definition of social media slop. just crank it out for ad vews and revenue, crank your schedule, crank it out, crank crank crank, nobody cares about quality or impact or art or meaning, just crank dat fraufraud revenue boi. You have no intent, no art, no purpose. You’re just cranking out dead internet garbage for some interaction criteria to manipulate other people. also see: AI image or AI video model generated false-human subscription content or AI human model, AI image, AI video generated political propaganda and other “no care from me, intensity for thee” outputs. “False creator effort” outputs are definitely slop because they’re made with the intention of manipulation or profit advantage for yourself when you have no direct personal creative involvement other than manipulating other people to some goal away from their natural instincts or inclinations and towards a manipulator’s mindset.
  • not slop: I would argue it is not slop if you are generating output, which both A, you do not take seriously, and B, you do not expect any other person or entity to ever take seriously as well. And C, you are not deriving any economic advantage from. So, D, there is no ulterior motive. You are only generating either examples, entertainment, or just abstract silliness evoked out into the world. Now, the one exception here is: even if you do not take something seriously, if other consumers do take this seriously or get dopamine loop addicted to your arbitrary output, then content could be considered slop in other people’s brains. Consumers of content end up sloppifying their subjective experience (generated by you, intentionally or not) via their un-mediated brain consumption, even though you did not intend to be creating any internally meaningful experiences in anybody other than a slight “heh” perhaps.
  • not slop: The most divisive category of not slop is content you have auto-generated, but then your creative process involves actually actively personally iterating on editing, refining, conforming, managing, testing, refactoring, and all other combinations of detailed analysis and growth of the content as a whole. Code can be auto-generated slop if you’re just blindly sending a request and blindly running the output without refactoring it or architecting it in the middle. But also, auto-generated code can be “not slop” if you generate it with an architecture and intent and security design and interaction design and observability design and management design, then you iterate and refactor on those goals until a coherent system is actually better than what you would have done yourself; you just created the results anywhere from 10x to 100x faster than doing it manually because you know, yourself, you can only write like 200 lines of new de novo working code per day, but the auto-generating system can make anywhere from 5,000 to 10,000 working lines of code per day if you’re actively monitoring the situation with caution, taste, and evaluation and reviews every step of the way.
  • the line: I personally would draw the slop versus not slop line around intent and exposure. If you are playing some “exploit the global big tech ad scam universe” game by making auto-generated AI images to post to social platforms which pay per click or pay-per-hit through having like shocking or unstable image things, instant slop, go to jail one million years. You can think of things like shrimp jesus or outlandish real estate pictures overwhelming low-information-consumer platforms where the platforms also have broken “pay per view” metrics so they seem to want infinitely generated dopamine slop as downward spiral forever platforms. So you have insincere output combined with intentional manipulation. This is clearly sloppy slop. But on the other hand, if you’re personally just making shrimp jesus because you want to be silly and fun, and you’re posting it to myfunsillyshrimpjesus.com with the intention of “going viral” and downstream causing the manipulation of thousands or millions of people into some circuitous feedback mechanism which benefits yourself economically or in a power structure directly, then does it really matter?

But we must consider the other axes of artifacts here, where we need to distribute an understanding throughout the world where currently some people understand this, but it is not widespread enough, where any visual or audio medium can no longer be trusted for having fidelity to reality due to the statistical representational continuation power of computer models at this point.

The sentiment here is upstream of what many people now realize is the dead internet theory: basically you cannot trust text content online because text content can be both automated and easily insincere, right? we know global-broadcast un-clique-filtered social media sites are basically just giant bot factories for millions of individual spam operations operating out of low income countries at this point, right?

propaganda

which brings us to the role of propaganda in online discourse. Propaganda is only effective if the mechanism and the medium of delivery is treated as sincere to the recipient (and one artifact of propaganda is the author knows their messaging is often corrupt too). We’ve been stuck in these patterns of online propaganda for 15+ years now. direct cross-national influence operations have targeted online false human representation of false beliefs for massive influence of the US electorate into preserving or into causing anti-national nationalistic interest for the benefit of third parties against the consuming party of the information itself (remember to protect your warm water ports, y’all).

But threading this back through the AI slop discussion, the primary question remaining: does AI slop enable propaganda at scale? using a recently previous definition: propaganda must have a recipient, and the recipient’s mimetic guards must be low enough where they are not realizing they are being propagandized too.

So, is AI slop propaganda? The answer is, of course, a big maybe. The sloppy slop scale depends on the memetic defenses of the recipient. Wgiven the current imbalance in world-model brain knowledge around the world and within stratified societies, we know the US electorate is elderly and also vulnerable to manipulation due to just natural human cognitive decline as age takes on. The elderly want to retain autonomy, while also losing the ability of discernment of right from wrong and reality from direct manipulation. So how do we essentially memetically vaccinate people or society as a whole into realizing you cannot trust any generated media you are not experiencing in a real-life situation, and even then, philosophically, we all only exist within our brain’s own presented hallucination of what our senses are receiving themselves. So trust reality only as far as you are sure you are not being currently drug manipulated, schizophrenic, hallucinating externally or outright delusional yourself.

vaccinate me bro

Let’s talk memetic vaccination.

With normal vaccinations, you just get a little prick, right? Little prick, and you’re done. With memetic vaccinations, You have to learn. You have to learn to such an extent, your dopaminergic reward circuitry flattens to boredom when you are presented with outlying data. So your reality conformance brain checkers do not diverge from a natural state.

vaccine yo self

it’s been a while since we’ve bulk posted media generation different auto-generated galleries or informational posters, so let’s do it again.

say hello to: Facetious/Propaganda

I was playing around with the recent AI video generation models and I found a way to make it 5x to 10x more efficient on a cost-serving basis for my own usage. So I spent a few hundred dollars to generate a thousand different videos, which would have cost something like $20k to $40k to generate at “market rate token” prices.

I don’t know about you, but everytime “ai generation” is talked about I can’t stop seeing the fast-talking, condescending, sarcastic voices of all the anti-AI influencers. Adam Conover is always hiding in your closet waiting to jump out and run his fast talking condescending whining about how bad AI is shtick across the entire universe. I will have you know these generated video artifacts I created only used the equivalent power of about two space heaters running for a week. Also, people love to constantly talk about “AI water use,” and I’ll have you know I only took one shower during the week while these ran, and I only drank two bottles of water. That’s your AI water use.

Overall, the goal of this Facetious/Propaganda project is to see what’s in the current public models too. This model is equivalent or better than what OpenAI had called Sora previously, which they shut down because these things are painfully inefficient to run at scale. Even at my micro-personal-scale, I got it to work great for my own usage, but I can’t imagine trying to make millions of users control these generations at a high enough quality where you’re not constantly being made fun of and called out when your technology doesn’t work because the faster you make it, the worse it gets.

As for the Facetious/Propaganda and memetic vaccination part, I tried to generate multiplicative combinations of many of the same prompts or punchlines in different scenarios, so you can see how the models are representing different entities across their representational space, which can allow you to both learn you can’t trust anything online (so many people “one idea bind forever” when technology is introdued and weird/bad then never update their ideas ever again. No, “ai images” are not all having 10 fingers; no they are not always all shades of orange tint; no the people are not always glossy and shiny, etc). Computational statistical representational power of universal reality will only get better and faster and higher quality over time.

Watch it all at Facetious/Propaganda now. Inoculate yourself into not believing anything you see in a digitally captured medium, especially if the goal is to somehow influence or manipulate or give you a one or more of what people tend to call serotonin boosts or dopamine hits or anything making you want to keep having more of something. If you consume media content online and it makes you feel “shocked” or “engaged” with attention spike editing, you must assume it is not real from now on (except for ones actually real, but then you’re told they “are fake” when they are actual war crimes or powerful people caught in embarrassing or fraud or criminal situations or something. But the pixel by pixel difference will be indistinguishable from a false reality where people are propagandizing their enemies into negative lights versus the actual world where the bad people are being recorded doing things in negative lights and having negative judgment thrust upon them. So, where is the balance? The balance is the provenance of the source of the digital creations, which in itself is an entire can of rainbow-colored unicorn worms, we do not have time to spin through the electric can opener and unseal the signs and begin ingesting at this juncture).

conclusion

hire me to create your own ai propaganda videos at i do it good and fast and Facetious/Propaganda is only the beginning