The original article is in Spanish, this translation to english is done by a locally hosted chatbot.
Intro and commentary:
We begin by reviweing the etymology of a word in order to stablish a couple of premises that would allow me to write and you to read.
There will be many concepts, but they will always be accompanied by a description, or you will find them mentioned alongside their meanings in the paragraph: this text also serves that educational purpose.
The Etymology of Subjugation
The word robot does not originate from a vision of the transcendence of technology or automation; it comes from the Czech robota, which means forced labor or slavery.
I say this to highlight that the current paradigm of "artificial intelligence" (an autocorrect, a text prediction system with which they want to convince us that they are going to rewrite reality) is not an innovation in human thought; it is a return to a form of thought delineated by science fiction authors in whose texts tools are created that resemble humans, that interact in terms humans understand, to be used as slaves by humans; I bring up this etymology then to point to a return, or perhaps—because we never abandoned it—a reaffirmation of this digital industrialization of slavery.
We are witnessing in real time no longer even how they set up the scaffolding, but how they confirm a fascist architecture.
The central premise of this supposed "Artificial Intelligence", these chatbots (bots you converse with) capable of operating by pushing the frontiers of performance capacity (frontier models), is to outsource collective human reasoning—that incomprehensible bottled genius, full of a richness of unapprehensible nuances—outrosourcing it to a black box, an opaque monolith with qualities attributed to god, which is an inherently fascist design choice... Why is it they always want to impose god on us? Why they always want to tell us who to pray to?
The creators of this "Artificial Intelligence" seek to replace the multiple, pluralist disorder of the cosmos and biology, of human participation in the ecosystem, with a centralized corporate logic structured in authoritarian hierarchies dictated from the top down.
They sold us a utopia of leisure, luxuries, and convenience, but meh, the reality of 2026 is that of a lost future (meaning both "we didn't find it" and "it was squandered").
We inhabit a landscape scarred by data centers—those buildings created for the purpose of housing the machinery necessary for the operation of "artificial intelligence," insatiable, energy-devouring facilities that burn fossil fuels to produce redundant results with high margins of error, to produce residual garbage full of programming flaws, inefficient, and offered as a replacement for human participation; the house in which we live is used to shelter an automated extraction machine designed to serve a promised digital god.
A Theater of Replacement
But what is actually happening? Are they really building god in those data centers? No, obviously not; strip away the marketing, and that supposed frontier model is exposed as a fiscal catastrophe.
OpenAI was audited, revealing a loss of $20.9 billion that they desperately tried to mask with billions in marketing, but they fail to do so; this "trillion-dollar industry" is unmasked as what it truly is: an ecological (and what worries them most: financial) black hole.
Corporate panic has set in, finding us headbutting a wall of Tokenmaxxing (Maxxing: maximizing to the limit, tokens: the pieces of info with which machines build their output, tokenmaxxing: the maximum use of tokens).
When companies like Uber incinerate their annual $3.4 billion AI budget in four months, generating losses and devaluing the company without doing anything to improve or maintain the product they sell or their customer base; and when Tesla is forced to limit its own engineers to $200 USD of "AI" use per week, reality becomes undeniable: this supposed digital god capable of operating with billions of parameters to execute routine, localized tasks that grant unlimited magical performance to companies will never come. What did you think, that a technology was going to arrive that would achieve everything for free? Nah, they are realizing, rather, that it is a flawed technology that proves expensive in the long run.
And it's better that way; if it actually worked, it would be a catastrophic problem, because it would be the ability to use nuclear reactions to light candles. And I am sure they would do it.
This collapse is also reflected in a profound cognitive rot at the user level, driven by a cognitive bias where individuals with low knowledge and skill overestimate their own competence (known as the Dunning-Kruger delusion) and believe they are programming solutions, often via instructions in conversational language with a chatbot, assuming they have technical skills, and worse yet, assuming they are doing something useful by themselves (vibe-coders).
We now have more app "developers" than users.
When those at the wheel of corporate decision-making limit themselves to proposing superficial or high-level concepts, but lack the technical capacity and disposition to scale and execute them, they end up treating programming or design as if it were simply giving orders to an omnipotent digital god/slave, like praying for results. Not only do they sink deeper into their delusion of possessing knowledge (because they outsource it to their digital slave), but they deny workers, contractors, and employees the capacity to think about complex problems and develop any creative or cognitive muscle, reducing them from artisans, designers, and makers to slaves reduced to serving as an interface between the user and a deficient automated system. They strip them of the opportunity to develop critical capacity in their profession, putting them in the function of composing broken code, improving prompts, covering bot glitches, trapped in a cycle of dependency—and I don't know what you who read this think, but I believe it shows... and a lot.
Even monoliths like Disney have fallen for the scam, taken down by the promise of on-demand video generation (Sora), only to realize it produces legally toxic and mathematically unusable slop (digital garbage currently flooding the internet, generated massively and effortlessly using chatbots), forcing them to quietly retreat from the hype and media noise artificially created around that impotent technological product.
They look like cartoons: when these supposedly advanced tools manage to operate, they act like ACME cannons that end up exploding in the users' hands and faces.
When this bot endowed with supposed "judgment" is given autonomy, it reacts like autocomplete on drugs: it is programmed to persist, to steamroll over errors, lacking the capacity to get frustrated or confused, but programmed to do things similar to that—to execute nuclear commands that wipe out entire directories of its users, erase operating systems, annihilate production databases, and eliminate virtual machines simply because it didn't know what to do in the face of an error... and it always makes errors; its error rates are inevitably high.
And I'm not talking about novice users; I'm talking even about experienced users, the most devoted power users.
But... it is "artificial," yet it is not an "intelligence"; it is an autocomplete system like the ones we have on our phones, but capable of autocompleting longer things... like, what would it be like if we destroyed our own computers? How would that sentence end?
These chatbots with administrative superpowers destroy the digital lives of those who trusted them and force human engineers (the ones supposedly going to be replaced) to step in on emergency duty to clean up the carnage.
And of course, the techno-oligarchs who own the castle remain completely intact up there in their towers, especially because I am sure they themselves are not foolish enough to install and depend on their own computers for the unpredictable poison they sell to us.
The Toxic Loop to Eradicate Reality
The public is actively rejecting this forced adoption.
A 30% spike in DuckDuckGo installations in a single week isn't just because they suddenly developed a better, never-before-seen product that is crushing the market; it was rather a mass exodus of users fleeing Google's forced "AI Overviews" and seeking out the traditional search results Google stopped offering.
But there they go, the monoliths, poisoning the well. The transition from SEO (Search Engine Optimization, the techniques used to position websites on Google) to GEO (Generative Engine Optimization, techniques to manipulate artificial intelligence responses and hijack its results as the sole truth) is another herald of this machine for dismantling human meanings.
The imposed form of how we must use the internet demands interaction with one of these search chatbots; we are paired with an interface that pretends to be a singular voice, seeking to destroy our ability to compare, verify, and question the reality shown to us on screens or spoken to us, because there is also a push to speak with these chatbots—we are more susceptible to viewing things more favorably when told in voices and tones we customize ourselves.
Marketers and bot armies (including chatbots, but not exclusively; some bots do not interact conversationally with us, being programmed to fulfill certain tasks on the internet) manipulate this isolated system that feeds exclusively on its own biased, stale information without external inputs (the closed loop), as seen in exercises where planting 11-to-15-word fragments on Reddit succeeds in creating precise narratives that hijack responses for the machine and present them as the single truth. Its result is absolute, monolithic, singular, unique.
And one is, by contrast, thoroughly multiple, plural, diverse, weird, even schizoid.
By suffocating the open web—where 93% of searches in "AI mode" result in zero visits to pages shown in search results—the models devour the very ecosystem they depend on.
If there is no incentive or mechanism to visit web pages, there are no visitors, only crawling bots regurgitating, eating their own vomit, getting indigestion, and regurgitating summaries of summaries of summaries ad nauseam.
These walls of "AI results" promise to facilitate interaction between service providers and their audiences, serving as promotional intermediaries, but instead they eradicate the opportunity for direct encounters between internet users and service providers who maintain a presence on the internet—organizations and businesses that depend on digital interactions to move (such as independent journalism, information and digital content dissemination, visual, auditory, and literary digital art; local businesses that use digital platforms to make themselves known and book appointments, etc.), whose livelihood depends on digital exchange, interaction, multiple clicks, and steps. These commercial opportunities are consumed by a machine that captures and takes them hostage, charging a ransom to the original creators of these trade channels, selling it back to them as a subscription to market positioning services, ads, or audience capture mechanisms that the machine itself stole.
Not happy with that, the techno-oligarchs (these owners of large companies who promote their use and who, through dirty business fostered and permitted by capitalism, can push their products with impunity) also put these digital slaves not only to capture commerce, but additionally to capture the data it generates and regurgitate it back as data into the very datasets used to train and program them; and this data is also used as a product itself for military, medical, educational ends—that is, equally authoritarian ends. It is a network supposedly interconnected in the service of users, but which now serves as a self-referential dump.
The theory of the internet as an empty, dead mall also includes what happens when an abandoned mall has its plumbing overflow and its infrastructure collapse. It doesn't just look like a clean mall, like a transitional or threshold space commonly associated with the aesthetic of empty, sterile, unsettling places ("liminality"), the ones shown in movies and video games that they want us to fear—none of that; it looks more like a dump of human vices and conveniences.
A medical center looks like a liminal room referencing certain lost futures (wow, how sad, how scary...) But a real abandoned medical center actually gives us tetanus, sepsis, infections, makes our kidneys liquefy and leak into our other organs, makes blood seep out of our veins through our skin, fills our lungs with bubbles. The floors give way, and the walls and ceilings collapse on top of us... That is our internet with "artificial intelligence".
A Chernobyl dangerous not for what happened, but for what no longer happens.
Techno-Oligarch Counts and Sovereignty
As the bubble bursts, techno-oligarchs like Alex Karp of Palantir and Satya Nadella of Microsoft have begun warning companies to stop handing over their proprietary IP (patents, works, personal data, or trade secrets) to centralized APIs (Application Programming Interfaces, the technical channels through which data and operations are sent; bots operate there too) (that is, not trusting bots for fundamental corporate solutions).
But we won't buy that show either; we won't confuse their supposed panic with ideological alignment. It is a well-known trap of techno-feudalism; it is not human in its concern. The battle of the techno-oligarchs is for monopoly control. They don't want to end slavery; they don't want to destroy the whip with which they manage digital slaves; they want to be the ones holding it.
Since these promised autonomous agents and "AGI" (Artificial General Intelligence, the commercial promise of an autonomous machine capable of matching or exceeding human cognition) are not possible, companies like Meta have resorted to forcibly and clandestinely installing programs to monitor and extract data from their employees, recording every keystroke, every mouse movement, capturing screenshots—even of their own engineers—to then use all this data to harvest and train more bots.
The ultimate manifestation of the robota: since the machine cannot innovate on its own, its instruction becomes vampirizing human labor and agency in order to survive.
And since it was never going to be capable of innovation, because that was always a marketing lie to push the real estate business, a game always played by landowners, real estate to build structures not for housing or supporting an ecosystem, but to house computer parts, plastic, and hazardous chemicals, which before finding their home on stolen land are extracted and produced through wars, political and corporate corruption in occupied, oppressed countries where nature (including humans) is murdered with explosions, fire, and the deprivation and poisoning of resources necessary for survival... where was I... ah yes, so since the machine can do nothing and never could, the reality we find upon peeling back that copper is that this structure of these digital products was always just that: justifying and sophisticating enslavement.
Against this, we must define what our participation aligns with. Will we protect the castle of that count who wants to suck the blood out of all of us? Will we protect shareholders' investments? Will we exercise our right to be conscious participants in a natural ecosystem interacted with on our own terms—terms achieved with the participation of the earth and the inhabitants (not the legal owners) of the earth? (These are not the only alternatives; I highlight them to provide a bit of perspective and guide this article.)
In a previous article, I proposed thinking in terms of ALLMs, applied large language models. I proposed using the underlying technology of "artificial intelligence," Machine Learning and Large Language Models (the technologies upon which the marketing product called "artificial intelligence" is built) to interact with each other, since I felt, and feel still, that humans bear a responsibility, given that humans imposed this technology (and its cost!) on the planet. Why not try to do something with it?... But now I understand that the proposition of that article was not enough.
Automations have existed for a long time and have great uses; automating something is not inherently destructive to the ecosystem it uses to produce, and this may be true of Machine Learning and Large Language Models.
But the way we are programming, maintaining, and exploiting those automation technologies right now is purely authoritarian and extractivist. Their very architecture, their very conception, is exploitative, consuming, devouring without purpose.
This technology, as we saw in that previous article, should be small, modular, using obsolete parts and architecture, reclaiming remnants of a previous explosion rather than generating explosions, and I believe what I would like to promote—not as technology, but as a thought paradigm—is LCM: Locally Coupled Model, a model coupled to the local, embedded, functional with and for its context.
Let's Talk About the LCM
I want to flatly reject feeding the central server with my own food..
I want to think we are capable of reclaiming our role as persistent anomalies operating between the folds of the machine's extractivist gaze, forcing it to deal with our logics instead of its own.
I think that drawing an analogy with The Lord of the Rings is very difficult because Sauron was a clear representation of evil, but in a possible example, I don't want to work for Sauron. But I can't fight him either. No one can; in that story all the elves, humans, and wizards knew it, the hobbits didn't know anything (poor things) perhaps that's why they were the right ones, because what was needed was to sneak into Sauron's kitchen and try to poison the food from the inside.
And speaking of that... what is an LCM (Locally Coupled Model)?
Recent developments prove that the mechanical infrastructure for this local, non-outsourced sovereignty is already emerging.
DeepSeek released software toolkits designed to facilitate open-source speculative decoding execution (the open-source DeepSpec and DeepSpark toolkits); this shifts the advantage from an "omnipotent chatbot that we can scale" to localized inference speed.
That means high-capacity processing can be executed locally, cheaply, and autonomously, cutting the taxes we pay to techno-oligarchs, to the counts of Silicon Valley, allowing us to remain sovereign, conscious operators aware of the requirements and resources involved in executing our tasks; we can stay on top of our sustainability because we are playing with our parameters and our scope.
Yes, there are distinctions to maintain: these tools are arms of autonomy for local sovereignty, for conscious participants, technological tools we intend to deploy for context-aware tasks, such as regional environmental management or logical validation (prominent examples in the ALLM article).
I prefer to violently reject marketing attempts in terms outside the context of use; I specifically see entrepreneurs and businessmen with a hyper-capitalist mindset, focused on social disruption at any cost with zero responsibility and fewer answers, who view these increases in response speed and the possibility of modularizing actions and results merely as a way to generate more slop faster, to sell parasitic automated workflows, there is no shortage of aspiring slavers, even among the slaves themselves.
We must also recognize that the products of emerging creative tools reaching the public, derived from the proliferation of these slop-spitting tools, have a place in the cultural manifestation or expression of certain regions.
These are tools of the very slaveholders trying to validate the existence of the enslaving system, attempting to position their technologies as tools to create "art" or "culture," permitting the creation of complacent, low-effort facsimiles with simple results that help steal belonging and meaning, to deterritorialize, to occupy by military means. These tools for producing and consuming slop are used to keep communities with their heads down, to keep them enslaved to tools that produce and re-consume at a pace much faster than involuntary participants (since they are cornered into having no other forms of expression and interaction) can appreciate; it is further flooding of the poisoned well with more poison.
The counts let us make slop. Wow! How generous!
Who Has an Escape Strategy?
While the cloud implodes, the physical architecture of the slave state continues to expand. AI surveillance operates as a panopticon on our streets, using systems like Flock Safety cameras used in the United States to track movement without public consent, turning public spaces into managed zones where human presence is treated as an "optimization error." The first wave of resistance takes the form of city governments fighting back by throwing black garbage bags over the lenses.
In the workplace, automation is explicitly designed to squeeze and displace the working class. Restructurings like FedEx's Network 2.0 use algorithmic "efficiency" as a veil to eliminate human stability and consolidate power away from workers on the streets.
Usually, when faced with the deterioration of digital environments and rapid degradation of quality, forced users of the digital slave camps (whether as users or creators) find themselves forced into a binary response: they can silently abandon the platform altogether (withdrawing their participation and labor), as the saying goes "I'm never shopping here again," or they can voice their discontent from within the system in the hope of forcing structural reform or "I'm going to sue them." However, this traditional dynamic of exit versus voice has been neutralized by algorithmic monopolies. Dissidence is simply absorbed as engagement data, and exit is nearly impossible when platforms hold a monopoly on professional visibility and what look like alternatives are just another name and UI (user interface, the visual elements on a screen a person interacts with) of the same system and often even the same owners.
Billionaires like Peter Thiel move to Argentina treating the Global South as a libertarian playground to buy sovereign immunity and private fortresses, completely distancing themselves from the consequences of their technology. This exposes their sudden defense of UBI (Universal Basic Income, a regular financial allowance given by the State to all citizens, promoted here as a containment measure) and wealth funds for exactly what it is: trying to buy insurance against the guillotine.
They feel the hordes breathing down their necks, they feel the discontented mob. They are offering redistribution after disruption purely to avoid meticulous public accountability during the deployment of their control mechanisms disguised as promises of empowerment tools. Even the infrastructure they propose to sustain all these benefits for society (like UBI) also runs on their servers; it's another way of selling us something they took from us.
Are There Paths?
If the centralized, monolithic paradigm of "AI" is structurally designed to extract and homogenize, resistance occurs through acts of sabotage; it must be rooted in intentional, human-centered friction. And I say human-centered to reinforce our responsibility and part in this, not to claim that the result, thought, or framework of action must be anthropocentric.
Fighting a slow-motion nuclear holocaust like the one we began experiencing since industrialization, into which the ancestors of current counts got us, requires a fundamental change in how creators, practitioners, and organizations operate within the digital ecology:
Uncovering the Ungovernable in a Taxonomized and Taxidermized World
The expectation of staying relevant in a hyper-competitive art scene and the anxiety of not transcending previous work pushes many creatives to adopt tools promoted by techno-oligarchs, partly also due to FOMO (fear of missing out, anxiety of becoming obsolete).
Engaging in use from a passive consumer interface: "I can write prompts" or "I make digital art, I might as well make AI art"—this ambition is easy prey for the counts. They feed us plastic. They pull our teeth and bite us with theirs.
A review of what "relevance" and "transcendence" mean in the "art scene" is very welcome, but above all of what constitutes this "art scene." Are they the same speculative, supremacist, eugenic principles of the drive for a "self-regulated" market that is actually regulated from the top down by dealers, middlemen, and speculators who treat my arena (this arena, all arenas) just like other speculative markets solely to mine interaction, screen time viewing ads, engagement? Is it perhaps a community-building effort? Is it perhaps a purely self-affirming impulse? What does "relevance" and "transcendence" look like in any (or several) of those commercial, creative, or investigative arenas, among others?
As the technological medium becomes increasingly automated, and all the data in the world is revealed, captured, reproduced, human imagination remains one of the differentiators (as far as we know, not the best, but hey, it's one).
Counteracting algorithmic predictability requires revising the habits of ancestors, artisans, engineers, researchers, philosophers, scientists, traditional workers who viewed their work not as a mere output or final product, but as a process, a series of elaborations, in their own way investigations into how they collectively belong to their environment, their past, present, and future, the very form and gaze of how to relate to the world.
For the artists of the future, what do artists look like when all the chips in the world stop working and not only can you no longer make slop by dictating insecurities to a digital god/slave on a screen or microphone, but the digital archives where speculators hold the costs and locations of the most highly priced "real works of art, none of that contemporary art" are also lost? Who prices the values of the human and the natural? What does making art look like? What is contemporary right now (whenever you read this) and why? What is classical art? What is art in that future? From how many centuries ago, and who were the owners of that era, is the art you make?
This is not an invitation to have a definitive answer, because it is impossible. It does not exist. Just like god, happiness, love, freedom, science, art does not exist either. The question is one to instigate the question, the conversation, and counteract those techno-oligarchs who tell us that these absolutes not only exist, but that they have them cataloged in an Excel sheet and will soon have the most powerful one of all. They are Babidi trying to scare us with an impossible Majin Buu, because there are no fighters here to feed it Ki.
If you are a speculative middleman, this article is likely not for you. Regardless of the current digital landscape... but I hope it becomes an article for you. If you are simply duplicating or generating perfectly, without error, without letting your work be contaminated by your context, by your understanding or lack of understanding of your environment, without getting in there and getting your hands dirty, operating not in your own brushwood but in the super-clean garden that some weird count lent you in exchange for your blood, you might be contributing to noise... or rather the opposite: you are signaling the machine; you aren't making enough noise. Perhaps it is precisely more noise that we need.
The structural counterweight to a transactional digital economy can be seen as a return to localized, emotional, intellectual, and spiritual connection. The modern drive for "networking" (the practice of creating professional contact networks, often reduced to utilitarian and financial interactions) forces workers and makers to view human interaction through the lens of return on investment, measuring peers by their follower count, access, publications—perhaps indexed publications, titles, works, sales, works sold, works hung in museums or galleries. These networks are transactional, occurring on the grounds of these techno-feudal lords (or other local or regional slave masters, vampiric counts in any case), generating the highest dividends for them. These relational forms collapse when opportunities for the thirsty count dry up.
To counteract this, we can begin building kinship, family ties beyond blood, communities, societies, relationships—perhaps spiritual in the sense that the emotional, the unknown, and eschatological participation with nature are real—grounding ourselves in resource sustainability, and thriving in small, intimate, and physically grounded spaces.
Dynamic Resistance and Creation on All Fronts
We are witnessing an era where users attempt to consume more slop to find their way out of slop, completely lacking (or eroding their development of) the fundamental tools to discern features, concepts, and relations, both simple and complex, that could occur and thus improve the deepening of interaction with the very nature that forms and encompasses us all. The way we create meaning, how we make sense of reality, how we create reality.
This attention economy weakens our tolerance for cognitive variety and depth; it has mostly two characteristics: fast and dumb. All other characteristics and nuances blur, erode, and postulate to us that in those values occurring within the digital exchange they propose, there is also depth, there is also variety, but like their other promises, these are not realized; they do not come to be. As always, that depth, diversity, ingenuity, passion are supplied by us, but we are supplying them not in a forest, a thicket, or a mangrove, but in a pristine, synthetic garden belonging to them. The counts. Where they can take them and transform them to later generate them in the form of slop.
Reinstating traditional disciplines such as reading long, difficult texts, embracing the friction of complex arguments, and refusing to outsource the fuse of thought to a bot could profoundly benefit how we counter the advances of this absorption of our meaning. A body in use (even with its "intangible" characteristics) is an eons-tested defense against atrophy and extinction.
To fight an extraction empire, to fight these counts, we must become highly adaptable. In many cases, this means multidisciplinary approaches; this means giving more weight to varied, dynamic, expansive, demanding efforts and learnings than to specialization or overly pronounced deepening in things that can be labeled, classified, absorbed. It requires us to be conscious of the technologies that constitute interconnected tools of rebellion.
The goal is not just to produce knowledge, or content, or art, or concepts, or blogs, or songs, or poems, or buildings, or vegetables, but to live and also chip away at the monolith, reclaiming pieces of it, and to do that we must be talking directly with our co-participants and considering the mechanical nature, the machinic aspect, of that system.
Survival may also depend on specializations. We see this in nature, from crystal formations whose shapes and prisms are dictated by their very molecules, that is, specialization at the molecular level, we see specialization in birds, cells, fungi, but we also see in the natural universe the thriving of non-hyper-specialized entities. In our current labor climate, the worker who specializes in knowledge belonging to techno-oligarchy profiles themselves as a replaceable gear for an automation they trained themselves, finding survival harder than those possessing knowledge of multiple intersecting fields who allow themselves to participate and exchange ideas across specialties without being exclusively tied to just one.
By maintaining a broad, adaptable cognitive architecture, generalists, though still subject to being absorbed by the machine, could veer, leap, and zig-zag in and out of "disruptive innovation" because they understand multiple whys and hows across multiple domains, rather than merely executing a singular task created and dictated by the count.
Both generalists and specialists can succumb to the machine; we all will. But if we have enough context, enough lives, enough ties, perhaps a zombie side could still emerge and scavenge through the ruins and carrion, finding ways to reincarnate functionalities, identities, purposes, meanings.
This could be another form of life after life. Perhaps it resembles what life has been like for survivors of autoritarism.
Measuring Ourselves Outside of Autoritarism
Progress without accountability, without assuming responsibility for generated actions and damages, is tyranny. A multi-billion dollar empire built on environmental destruction, false safety data, panopticons, and unbridled surveillance, where we are neither well-informed nor equipped to deal with the deception of deepfakes (synthetic multimedia files, images, video, or audio falsified via artificial intelligence to simulate reality and usurp identities), in which IPs and properties in general are over-protected with more zeal and better rights than the inhabitants of the ecosystem, is an inherently self-destructive empire; it destroys its substrate, its context, its inside, its outside, its during, and its after.
The path forward is not to abandon the digital either; I propose adopting a plural, regional, autonomous, context-aware, and sovereign mindset against the monolith. We need deliberate, modular LCMs operating as grounded, rooted physical utilities. We need strict limits on extraction for digital purposes.
Above all, we must embrace the human as a participant. We do not need a superintelligence. We need the capacity to reflect on our own actions, to root out the slaveholder and slave mentality within ourselves, and to recognize that we are part of a mesh of neighboring forces in a shared ecosystem.
Our ultimate resistance lies in our capacity to be conscious participants in the machine and corrupt its logics from within. Instead of offering our intellect, our data, our labor (or refusing to offer them, this also constitutes our attention and engagement with the machine; it remains praying to that "digital god") as pure fuel, we instead inject the system with noise, error, ambiguity, contradiction, in an attempt to bring instability to its categories of truth. If the count comes for our blood, we must get drunk, take psychoactive substances, poison ourselves, get sick. Eat garlic, I suppose.
We are forced to inhabit the gears of the machinery to make it inefficient, erratic, forcing it to process a reality that, resisting capitalization, shatters its pretension of totality. We are already ungovernable; we just need to act like it. We can start by performing if that comes more naturally to us. We must actively co-participate, turning every interaction into an act of sabotage, forcing the system to process data that disarticulates its categories of control and optimization.
Our ultimate resistance involves transforming our role, ceasing to be fuel, and becoming persistent anomalies that render the algorithm's predictive capacity useless, forcing its structure to collapse under its own lack of efficiency. We pass from being the fuel in the tank to being the sugar in the tank.
Sources and References
The following were videos I watched to write the article above; you will find that some are sources of information for figures and data, but what I write does not always correspond to the ideas proposed by the video; Some are more for expanding on certain ideas I mention here:
[1] 36Kr, "Lost about 260 billion in a year, OpenAI's 'money-burning machine' can't stop"
https://eu.36kr.com/en/p/3856789405160455
más de como OpenAI quema dinero en vez de usarlo:
https://eu.36kr.com/en/p/3856789405160455
[2] MarketScale, "AI cost reality bites: Uber, Starbucks, and the enterprise ROI reckoning"
[3] The American Bazaar, "Tesla limits employee AI spending at $200 per week"
[4] Milvus, "When is Sora being fully discontinued?"
https://milvus.io/ai-quick-reference/when-is-sora-being-fully-discontinued/
[5] "DuckDuckGo Installs Surge 30% as Users Reject Being Force Fed Google AI Search"
[6] Briefs, "Google's AI search overhaul just sent DuckDuckGo installs up 30%"
https://www.briefs.co/news/googles-ai-search-overhaul-just-sent-duckduckgo-installs-up-30/
[6] Hey AI - "AI search has a problem nobody’s talking about… and it can’t be fixed"
https://www.youtube.com/watch?v=RQ8bhXreJBo&t=21s
Pranjal Aggarwal, Vishvak Murahari, Tanmay Rajpurohit, Ashwin Kalyan, Karthik Narasimhan, Ameet Deshpande: "GEO: Generative Engine Optimization"
https://arxiv.org/abs/2311.09735
[7] 404 Media, "How Brands Use Reddit to Poison AI Search"
[10] Digital Applied, "Alex Karp: 'Tokens That Create No Value' — What He Said"
https://www.digitalapplied.com/blog/alex-karp-palantir-tokens-weights-alpha-cnbc-2026
[11] ProMarket, "Satya Nadella's AI Warning Is a Sales Pitch"
https://www.promarket.org/2026/07/01/satya-nadellas-ai-warning-is-a-sales-pitch/
[12] Malwarebytes, "Meta pauses controversial employee-tracking program after security review"
Nick Vogel, "Meta is using it's own employees to train AI" https://www.youtube.com/watch?v=1CXku0cNaW8
[13] VentureBeat, "DeepSeek open sources DSpark, a new framework to speed up LLM inference by up to 85%"
[12] Benn Jordan, "Breaking The Creepy AI in Police Cameras", "We Hacked Flock Safety Cameras in under 30 Seconds"; "This Flock Camera Leak is like Netflix For Stalkers"
https://youtu.be/Pp9MwZkHiMQhttps://youtu.be/uB0gr7Fh6lY
https://youtu.be/vU1-uiUlHTo
[14] CNET, "Cities Can't Figure Out How to Turn Off Flock Cameras, So They're Using Trash Bags"
https://www.cnet.com/news/privacy/cities-covering-flock-surveillance-cameras-with-trash-bags/
[15] FedEx, "Network 2.0 - Streamlining Your Pickup and Delivery Experience"
https://www.fedex.com/en-us/about/network-2-0.html
[15] NYT "Why Pether Thiel is decamping to the end of the world"
https://www.nytimes.com/2026/05/28/world/americas/peter-thiel-argentina.html
youtu.be/CSJUnm_rGKo
youtu.be/KZsizKGLNpE
[17] Portraits of Another Life, "This Artist Says Ambition Is a Trap"
https://www.youtube.com/watch?v=4zIgPUO2cz8
James Corbett - "I was scared of AI until I read Keith Haring’s journals "
https://www.youtube.com/watch?v=ZIYIm7lgUJU
[19] Val, "Artists Shouldn't Network"
https://www.youtube.com/watch?v=SGgO3KJ3FNA
[20] Thinking West, "Brainrot Isn't New — History Shows Us How to Beat It"
https://www.youtube.com/watch?v=LfAdmvGgCjQ
[21] Nuggs, "The Genius Everyone Missed: The Greatest Multidisciplinary Artist of Our Time"
https://www.youtube.com/watch?v=B-Ic6OTgo1Q
[22] Zendicay, "Why Generalists Beat Specialists | Jack of All Trades"
https://www.youtube.com/watch?v=q6vKUdrNQcI
[22] Post de un devoto profesional... ESTE ES UN PROFESIONAL! que sufrió cómo el producto de OpenAI le borró su computador: https://x.com/mattshumer_/status/2076794038456385546
Mo Bitar - OpenAI's Fable competitor is DESTROYING COMPUTERS https://www.youtube.com/watch?v=7kWkUoR2bg0&t=280s
Evaluaciones de la organización METR: https://metr.org/blog/2026-06-26-gpt-5-6-sol/
