Dark Tech speculation
This essay will explore a number of ways that technology might change in the future. This could be interpreted as a mix of research, art, and science fiction thinking. The goal is not to describe perfectly what might occur, but rather to provide imaginative exercises in what might happen, with the intention that researchers and thinkers might work to anticipate and build ideological antibodies to these kinds of post-capitalist, techno-feudalist concepts.
GPT meetings
In the first scenario, people involved in online meetings could use a GPT model to generate their agenda and action items. Recipient members of the meeting could use GPT models to listen to and summarize the meeting for them. The end result is effectively a number of GPT bots speaking with one another.
How to make money? The platform could provide membership benefits, including access to better models of the GPT technology to pay members. Non-paying Members can receive access to the basic model, which they will not be confident is not hallucinating, pulling bad data, or otherwise faulty or inferior in some way. The platform can also systematically weaken or make obsolete the free models over time, degrading their quality.
The social impact is fairly evident: just as we judge each other for the authenticity of our leather shoes or price of our perfume, we will judge each other for the quality of our GPTs/LLMs. If I pay more, I can have my GPT synthetize and beautify my ideas for the meeting, perhaps even generate new and novel thoughts. My boss' GPT will send a memo to my GPT thanking it for the impressive ideas. We will all be bosses with robot secretaries, in a way, like ventriloquists whose dummies have come to life. Mass schizophrenia will break out, which of course will require GPT therapists.
Dating app MMORPG
In this scenario, people choose from a template and admixture of AI generated thoughts and ideas to design an avatar that dates online for themselves. The Avatar can operate multiple social media accounts, post on automatic scheduling, and develop a personality over time. It would resemble a cross between a digital pet toy of the 1990s, an AI assistant bond, and a sex worker as fantasy persona.
The appeal of these bots would be the same as one might find in an massive multiplayer online role playing game, or MMORPGs (what a word to say aloud). In dating others, persona can accrue points, credit, and obtain certain items. A body count scorecard could be utilized, to indicate how many persona the Persona has slept with digitally. The pleasure that human users would receive would be based on upgrading and leveling up their dating persona, until that Persona appeared more regularly at the top tiers of the dating app hierarchies that are programmed into the algorithm. Put another way, a high level character will swipe less than others and be swiped on more often than others. A swipe ratio will be one way to assess level.
How to make money? Subscription plans could offer you basic access, with higher sub tiers offering broad visibility (time and space) of your personas placement in the algorithm, as well as back end data metrics about when were the best times to post and which of your posts received the most reception. This occurs already on most influencer-based platforms. Third parties could sell digital extensions or overlays, as in the v-tuber industry. While the ultimate goal might be to really find a partner (the assumption being that your persona's personality is close enough to your own that a match would make sense in real life), in reality the dating app will only function well if people return to it consistently. The fantasy that one is truly dating and 'winning' the dating game through a good ratio is the appeal. Actual facial, physical, and sexual content would start to feel old fashioned, complicated, and somewhat disgusting.
Work is stupid
Advertisements on platforms will increasingly encourage users to think of work, that is writing, reading, and critical thinking, as something fundamentally stupid and not worth their time. This will run counter to larger narratives about the increase of labor, longer working hours, and low wages. The ads will encourage users to devalue their own work and to think of their jobs as something that somebody else should be doing. It will emphasize the importance of all users becoming rich as soon as possible, through various uses of the platform tools and architecture.
These advertisements will be simple and straightforward, often with explicit statements and messages. At the same time as they are running, most people's jobs will feel inherently stupider, as GPT models take over much of the thinking portion of their work. The sophistication of the models will generate two outcomes: it will increase the quantity of total communication, necessitating additional bots to receive and process the output of the first bots. Then it will promote the idea that labor itself should not be a thinking or working process, but rather something that people should be freed from, not discover or manifest their skills within. All labor will be reconfigured as product, not process, and anything slower or less comprehensive than GPT output (i.e., anything humanly flawed) will be thought of as significantly deficient. Speed and accessibility will be prioritized over depth and complexity.
As such, luxury sports and amusement, such as skiing, scuba diving, cycling, video games, and any sort of Avatar work, will become increasingly popular and sold via these advertisements. The concept of play as work will proliferate, and people will be encouraged to believe they are only being productive when they are leveraging their emotions in a powerful and potent way. The games and sports they engage will largely be virtual and data-tracked in order to produce additional advertisements. In earning less, working less, and consuming more, the average consumer will accrue enormous amounts of debt and eventually be forced into fields of manual labor devoted to the maintenance of the large-scale server farms that power GPTs. This might be cleaning, clearing land, physical maintenance of hardware, or building shelters from superstorms that threaten the computes.
Therapy gig work
As emotion becomes the primary output of human laborers, many people will begin to operate and think of themselves as amateur and necessary therapists for their peers and friends. While GPT models will be able to provide some level of therapeutic relief for users, these models will be locked behind paywalls, their sophistication degraded or fragmented across paid hierarchal tears, to ensure that subscription models underpin the deployment and efficacy of these therapy bots. Furthermore, most bots will have general bans on providing therapy work, at least at low paid levels, to human users.
Simultaneously, gig workers will take up the mantle of therapists, using online therapy degrees they generated with bots at bot universities. There will be pressure to create a new social norm that only real humans can have real feelings, and that these feelings are worth paying for as a user who needs therapeutic relief. A range of services and types of therapists will appear. The first will be renting a friend for the day. The next would be renting an intimate partner for the day, this resembling traditional sex work, with the addition that the sex worker is understood to be part of therapy as well. As GPTs colonize spaces of human thought, emotions and sex will take on new meaning as the last bastions of pure humanity. Some will turn away from this, uncomfortable with their bodies and their feelings. Others will embrace it like a religion. Most people will fall somewhere in the middle, compartmentalizing animal instincts via sublimination methods of masturbation, extreme sport, or chemical addiction. What will truly be new is the role of sex worker therapists in society as icons of pure humanity, in contrast to the cold hardware of GPTs. Of course, GPTs will gradually colonize sexual spaces as well, via virtual pornography and personalized sex-talk work.
The human resources field will be transformed, and its name will take on new meaning. Gig therapists will be brought into dysfunctional offices or into dysfunctional families. Emotion being so valuable, and a flesh and blood practitioner being perceived as the authority on emotion, will mean that all kinds of work will emerge for good workers who can practice and hone skills at practical empathy, self-help and well-being, and other forms of deployed non-reflective, productive emotional labor.
Chatbots will race to keep up with these human therapy gig workers, and there will be an engaging, controversial push back against this. Whether Chatbots will become more human will stop being the main question. Instead, the question will be whether humans become more like chatbots. Sex therapists will embrace imperfection and physicality, offering what will quickly become small cults of pleasure and pain worship where social interactions become chaotic and inconsistent, as this is felt to be more 'human' than the programmed and safe discourse offered by chatbots. Chaos will once again drive cultish behaviors in a human society confused by disrupted technological change, displacement of previous moral centers, and degradation of social norms and communities.
Critical production
In this scenario, there are multiple phases.
The first is occurring now. GPT bots are currently producing a mass of reviews and critical receptions of the latest games and movies. They assemble text to maximize search engine optimization, and follow well-established templates to generate blog posts that review, analyze, critique, and offer vanilla, bland commentary on products.
As these sites proliferate, they will become less valuable over time. Human users will know what to expect to read, they'll learn the templates and vocabulary of the bots, and become basically bored reading the same content. This will hurt ad revenue on these sites, and a new model will be called for.
AI image generation will improve to the point where video games and movies, in simplistic form, can be generated on the spot. Blending the text and vernacular archives of movie and game reviews, a GPT model or neural network model sophisticated enough could write and generate a movie or video game on the spot, at first in variations for portions of the audience, then eventually personalized for each user. Personalized media will last for a little while before costs and environmental destruction makes it prohibitive. There will be, therefore, a brief 'golden period' in which everyone gets their own full-length entertainment galaxy.
Subscribers will be able to generate games and movies via prompts to the mod. For unpaid users, their video game or movie will include advertisements, as part of the unpaid subscription to its service. Paid human viewer subscribers will be able to appreciate the game and movie at full speed, optimize performance, and subscribed creators can create with a broader range of archival data and therefore greater variety in movie or game narrative and imagery. Subscription tiers will be thorough, omnipresent, and intricate. Price models will be varied by corporations using A/B testing, identifying classes of consumers they can charge more or less at different points. Prices will fluctuate fairly wildly as the economy destabilizes frequently in a hyper-capitalist flow of minority mass capital shifting between tax havens and trend stocks. Working people will see their wages likewise fluctuate as work demands change drastically week to week. Prices and wages will exist in an antagonistic twinship, like siblings punching each other in a fist fight. Entertainment will dominate people's free time, as it does now.
This will not end the human element in movie and game production. Instead, what will happen is human creative actors will employ technologies to develop their dream projects without the collaboration of other humans. As a result art will become deeply idiosyncratic and personal, and its form simultaneously unbelievably generic and bland, like what content creators do now. This will reflect the creative and imaginary resources of the human users who pay for neural networks powerful enough to produce their dream movies and games on demand. Talented creators will make odd art; mediocre creators will recreate popular ideas.
This leads to a further development: an end to human collaboration.
End of human collaboration
As human labor is considered foolish, and GPT models can produce much of what people perceive of as work, including creative work, the difficulty will be in human collaboration, that is team building, holding meetings, negotiating on ideas, arguing about outcomes, and being exposed to alternative lifestyles or worldviews. These activities will cease to be of substantial economic value, because they move too slowly compared to AI.
The end of collaboration will mean human psychic resources are increasingly focused on their own narcissistic worldviews and outcomes. Selfishness will be seen not as a vice or virtue, but rather as a default and rational mode of being, akin to ordering take out rather than cooking at home, because collaboration with others no longer becomes necessary.
As a result of these factors, human collaboration will displace itself from face to face or simultaneous, synchronous interaction, to displaced, asynchronous collaboration that occurs across time and space. People will collaborate, in other words, via memory of what others have said similar to the work of literary agents 500 years ago, who read books by authors far away from them in time and place. Creative actors and those humans seeking intellectual connection will interact with the GPT outcomes of their would-be peers, and produce their own replies in turn. They will be a kind of correspondence collaboration, rather than a real time interactive one. This is similar to what happens in the workplace when a group of GPTs hold meetings while their human users watch screens elsewhere.
In many cases this might lead to extensive, rich, and sophisticated collaboration that is asynchronous and form, much like letters between politicians 300 years ago, or odes in poetry. There will be a fundamental sense of yearning in this era. People will become highly skilled at talking about their own thoughts and feelings, and highly descriptive in discussing the thoughts and feelings of others, having to use their imagination of these thoughts and feelings. In alienating ourselves, we will retreat into our imagination of what others are like, and produce rich, fantastical, dangerous assumptions.
As a result much of the generated creative work will slow and individualize. Creative vernaculars will emerge out of pockets of thinkers who interact with each other voluntarily, out of similarities of taste rather than geography. Humans will never get off the planet, having developed technology that helps us to become more narcissistic, inward looking, and alienated. We will un-discover the world and begin a second 'dark age' of global life, in which countries shrink down to provincial territories fundamentally uninterested in talkin with one another. Only a minority will engage in distant correspondence, just like scribes and travelers centuries earlier, and from them will pour all kinds of metaphysical theory and guesswork, helped in no small part by the magic AI boxes they are bound to.
Fan citizenship
As people interact in this way, through highly self-referential creative outputs mediated by bots, citizenship will be based on fandom, that is attachment to fantastical and escapist products. Nationality will cease to have meaning; hedonistic and escapist fantasies will form the basis of social bonding. Digital communication will render much time and space essentially meaningless. What one likes, as a sincere expression of human emotional preference, will be more important than where one is or what one does (since most jobs will be the same, anyway). Class will disintegrate not into a utopian classless state, but a uninterrupted social desert of very few human users in a vast, bot-dominated internet filled almost to the brim with advertising. In this milieu, people's fandom will be channeled into purchases. It will be economically beneficial and thus socio-culturally acceptable to spend large percentages of one's earnings on toys, and then to share news of your purchases online, write amateur reviews, express a range of emotions centered on the toys, and react to others who do the same. Much cultural life will revolve around rituals consumption and desire, and not much else.
Amnesia and Nostalgia
Overwhelming amounts of data will create in users a curiosity about larger organizing structures, i.e. why am I seeing this?
Data flows will be algorithmically arranged and thus not subject to human criteria. As such, in the crisis of information glut, human viewers will define their choices along lines of nostalgia to grant an underlying and reassuring logic to the constant stream of information. Nostalgia will provide a narrative for people to shape and channel the firehouse rush of information coming at them: they will ignore information that contradicts their nostalgia, and embrace information that confirms it. This is more than political preference or bias; nostalgia is importantly a childish, sentimental framing of reality that sees all that is happening in the context of how it serves an earlier, purer happiness. In an era of hyper sophisticated machines that appear to know everything immediately, humans will fall back on sex and therapy as well as the imagined innocence and purity of childhood and infancy, in part because at the earliest stages of life humans do not have the capacity to know things, and are thus relieved of the pressure to compete with machines. We will infantilize ourselves as a way of escaping the demonstrable informational superiority of the machines. Whatever else they may have, machines will never be 'babies', will never be 'pure of heart', and will never have 'animal instincts' about life.
While the constancy of influx of new information eliminates the perceived need for memory, the desire for structure will be satisfied by a rough-edged emotional reaction of nostalgia. Thus, amnesia and nostalgia will operate together to repair the human psyche in its struggle against incessant information. The marriage of “long-memory” nostalgia to “short-memory” amnesia is not ironic, but suitable to the cognitive dissonant state of tech-enveloped human viewers. We will constantly forget everything that is new while deeply remembering (or inventing) everything that is old.
Institutional parasites, AI and universities
The exclusivity that defines elite institutions will break down as GPTs are used to infiltrate the informational corridors of hallowed halls. It will not be a secret what governments, experts, professors, and doctors think: the GPTs will know, and to users who pay enough, they will tell.
Institutional knowledge will furthermore be stolen and processed by GPTs which can then use them to undo institutions, start their own, or benefit competitors. A cheater at a small school can use a GPT to learn about education at a big school. Old gate keepers will lose their locks when they give their data over to the massive GPT databases. Even if users are barred from asking dangerous questions, the corporations themselves can force themselves into the elite by merely scanning the information they've stolen. Informational chaos will bring institutional ruin. New institutions will arise, but the line has been crossed: there will be no privacy in a world made to feed robots.
AI for family photos
In the future, cloud services for storage will be extremely expensive for many people, and the labor of taking family photos will be considered extreme. Instead of gathering family photo albums, people will prompt image generating AI bots to produce family photos for them, based on a repo of family member body scans typical of government and employer standards.
This is the list for now; perhaps if more occurs to me, I'll post again.