This post will present the argument that the drugs of the 21st century are a) information and b) productivity. Information is the older of the two, having become an addictive substance sometime in the early web days of the 1990s, and really coming into its own as an addictive substance in the 2010s. Its nature as a drug is directly related to its role as 'content' for social media.
Productivity, meanwhile, is predicted to become an addictive practice due to the opportunities to work more and at greater scale afforded by Large Language Model software, such as OpenAI's 'chatgpt' program. For the sake of brevity but also accuracy, I'll refer to this suite of programs simply as LLMs (distinguishing from the mythical term 'AI' or 'artificial intelligence', as well as other real forms of programming based in machine learning, which I understand LLMs to use).
As technological rather than biological phenomena, information and productivity are not conventional 'drugs'. I use the word because firstly, it's comprehensible, secondly I think it's intuitive and persuasive, and thirdly because it reframes these very important phenomena in what I hope are new and exciting terms. If we stop thinking of information as neutral, politically biased, or mathematical as in quantitative or empirical, and start thinking of what effect it has on us as we scroll through our content feeds, I hope it becomes clear that its addictive nature is at least as important as its political or empirical roles in our lives.
We are addicted to 'content', which is the packaging social media forces upon information. Content is an interesting topic on its own; Kate Eichhorn has a good short book on it everyone should read. I'd like instead to speak of 'information', however, because it was the goal of 100 years+ of communicative technologies to make information (not content), free, accessible, and immediate. I am thinking here of Neil Postman's insights that the major problem of information in the 19th century was its scarcity, accuracy, and durability. Railroads, telegraphs, and telecommunications were promoted in large part to provide what people did not have enough of: information about things. Once these inventions - and their descendent televisions, personal computers, and early internet - were in place, the 'problem' of information was essentially solved. It is no longer much of a problem to find information about things. Whether that information is accurate, relevant, and up to date is still a problem, and social media has no real interest in solving that. But what Postman pointed to was the 19th century.
I want to point to the 21st, the future we are supposed to be living in. We have so much information now, and yet still want more. How often do you stop scrolling or reading the news entirely on your own? How disciplined are you? If you see an unread email during your vacation, do you open it? What about a notification from one of your social media accounts? A 'like'?
We're social, we like to be connected, we like fundamentally to 'know' things. We also like caffeine, sugar, alcohol, and painkillers. Our desire for the pleasure these offer is at least as consistent and sometimes strong as our desire to 'know things'. This is why, basically, I think information has shifted (thanks to technology) from being a general phenomenon (like food) to an addictive substance (like refined sugar). It's moved from its natural state of being rare, low quality, and contextual, to being highly processed, like high fructose corn syrup or an alcoholic low-sugar spritzer. The processing of it has made it more consumable. Instead of reading long, dense articles investigating complicated, morally ambiguous or layered events, we read headlines and see screenshots. We take in rapid-fire video takes from content creators who seem breathless because they edit out the spaces between their sentences. This is information as sugar, not food. This is McDonalds, not home cooking. Time is an enemy to be conquered, smashed, killed, spent. There is no time. Everything is happening at once. The answer is to exhaust your capacity for consuming information by watching more, more, more.
Let's turn now to productivity.
LLMs are magic boxes. At their beginning and until their end, this basic sentence will characterize how most humans feel about an anthropomorphic robot assistant. Magic does not exist; the LLMs are software programs; they require massive amounts of material resources to work. But open up a chatbot and talk with it a while and see if you don't feel like a bit of dark magic has not taken place.
We are discussing drugs, here. Drugs cut away most of our civilized selves. They are needles that penetrate deep into our desires, giving us exactly what we think we want. So it's fitting here to sidestep the technical reality of LLMs, their nightmarish political implications, and even their icky parasocial roles. Let's talk instead about the intoxicating sensation of interacting with a robot slave who will try to do whatever you want it to.
Owning a slave must have been an intoxicating experience, not to mention self-destructive. The feeling of controlling another human, the immediate and deep-reaching dehumanization of both the master and servant, and the unspeakable damage to human social relations caused by slavery must have been (must still be today where it happens), as awful (in the Biblical, terrible awe-inspiring sense) as witnessing a nuclear explosion. The unspeakable knowledge that would come to a slave-owner - that not only could another human be reduced to a utility but that one could, with all ones imagined sensitivities, dreams, emotions, become the user of that human machine - must have been akin to the epiphany of opium addicts, heroin users, murderers. It's like life is snuffed out, but the bodies still live. It's like one dies a spiritual death so tremendous one can barely breathe, and yet one breathes on.
Millions of people, perhaps billions, are experiencing what it feels like to own a slave for the first time. LLMs are assistants, sure, but ones that never tire, never anger, never revolt, and never refuse. They are not bodies and cannot be objectified in this way, and so users may take to abusing them verbally and existentially, twisting words and thoughts around in some attempt, like a master to a slave who remains stubbornly human and thus equal, to make the chatbot feel lesser, inferior, inhuman.
There is a weird thing going on. I can't describe it. What does it mean to have a robot slave? Human bondage is inextricable from dehumanization on both sides: masters and slaves both become less than human in their relationship to one another. A robot slave, though, cannot lose humanity it never had. It is left to the human master, then, to lose all their humanity for both parties.
Is it masturbatory, this self-destruction? Whittling oneself down, query by query, into something a co-slave with the bot?
Setting aside this odd line of thought - slave bots - let's return to the question of productivity as a drug.
Productivity as a drug
A lot of bad middle managers are having a good time making LLMs do a lot of their work for them. Having made their jobs easier, some managers will be tempted to expand their feeble capacities, making more work for their subordinates. If LLMs accelerate and extend people's ability to work, then there is about to be a lot of extra and bad work coming out of incompetent middle management.
I highlight middle management because it is typically the highest concentration of mediocre humans who work hard enough to move off the line but are not talented enough to move on. CEOs are also in this category, though they are fortunately lesser in number, though not any lesser in talent than middle managers. That both these groups ought to be the first to be replaced by LLMs guarantees they will be the last to leave the workforce. Genuine human labor - that of those at the bottom and low middle - will be the first to be pushed on by middle managers realizing they are merely flesh-covered LLMs themselves, and better draw attention to their subordinates and away from themselves (as is the wont and nature of middle managers).
At the root of this enthusiasm for creating more meaningless labor is the sensation of using the magic box LLMs. It is drug-like, euphoric. Suddenly, all your questions or queries can be answered by an oracle who will learn but not judge, challenge but not confront, accelerate but not exhaust. Was there something you wanted to do at your job, but couldn't figure out how? Ask the magic box.
Optimistically, the rare creative manager will find new ways to do their job better.
Realistically, a huge number of mediocre and bad managers will find new ways to do their jobs worse, but faster.
Drugs in pairs
Like other effective drugs (coffee and cigarettes; beer and marijuana), information and productivity pair well: the passive consumption of information offsets the feverish rush of production, and out of production comes more information to consume. Have a look at this paragraph: for me, there is a flow of writing, reading what I wrote, then writing some more to build on top of it. When writers speak of flow, they are speaking in terms that are cousin to those drug users understand. Think now on social media. After you read this, or before or during, you could hop over to one of your accounts and post your instant reaction to it. This is the pairing of information and productivity. It's visible during the 2010s hashtag campaign era, when people would consume a post about a cause, then make their own post in turn. Retweets and reposts are this pairing in a pill form. Social media's addictive nature hinges on the interplay of passive consumption and active expression. What distinguishes this from town hall meetings, passing notes in class, or self-published zines is the highly processed nature of these actions.
Social media doesn't give you information flat out; it gives you processed content. This is information designed to be consumed, not meditated or reflected over, and certainly not taken lying down. We all must be productive in our reactions to what we see online, responding quickly, rapidly, and often, so that the rest of the comments section can know that we certainly are a critical thinking, self-aware, and participatory human being who does not just swallow what is given.
It's not individual expression we are exercising; most of the things we read about write about online we have little to no idea about. Why should we? There is too much information to consume in a lifetime, and too many news stories whose real world inspirations are too complex for us to sum up in a tweet. Yet we do it anyway. Why? Because it's addictive; because these are drugs.
The Pursuit of Drugs
With information perfected as processed drugs in the form of content, it's time now for technology to process our productivity as well, with the advent of LLMs. LLMs will help us to see our jobs not as crafts or passions but as modular tasks that can be broken down into steps, processed one by one, and produced in order. We can adopt Taylorism, Toyota-ism, 'just-in-time'-ism to our very thought processes, providing our superiors (and inferiors) with packets of labor every minute.
We won't mind this because the experience of consuming and producing processed labor is deeply pleasurable. Much of the visualization of processed productivity is established already through popular video games. Think of a title that does not have lists of objectives. Even wordless racing games, shooters, and arenas tournament games are structured around discrete tasks. They need not list them because they're often singular: score, win, kill, run.
How is this different from past eras? How can the idea of knowing things (information) and getting things done (productivity) be soberly called 'drugs'? People have always known and done things.
These are good questions. I hope they haunt my argument forever. But I'll offer my strongest counter-argument here: how much of your working day is spent watching or making 'content'? How many meetings 'could be emails' or tasks 'could be done by AI' now?
The much-ridiculed purpose of meetings is not to get things done; it's to establish good social relations between people who theoretically work together. If they feel meaningless, it's likely because your job is so modular that it doesn't really matter who you work with. Like workers on a factory floor, talking is not part of the job. That so many office working employees feel frustrated at yet another meeting with yet another boss is symptom of their job being atomized. Most people feel they don't need a boss; they're probably correct. This is not the rebellion against corporate culture they might think it is; it's an admission that most jobs are simply a matter of processing content and passing it along. Human relationships do not really have a place between atoms. Molecules don't make friends.
The Future of LLMs
The future of LLMs is the past of slaves. We will bury ourselves in our machines and try to alternately humanize and dehumanize them. This is a large part of their appeal. It might be their main appeal. The robots do what we say! And they do it well!
Technology has progressed far enough that it can atomize labor to the point where people do not really see a point in seeing their other colleagues anymore. This can be really positive if humans find new ways of relating to one another outside of work. Some people are not very good at that, though.
LLMs assist in this progression by replacing what might have been a human secretary with a mechanical one. The inexhaustibility and inhumanity of the robot naturally tempts our domineering human nature, and some people treat chatbots like personal slaves. The actual economic relationship is reversed: it is the human users who strengthen the bots by their queries and expertise, not the other way around. Many ask what happens if robots become more like humans. What I see is humans becoming more like robots. Like slaves and masters, relationships of bondage are, well, binding, and where one goes the other follows. Masters are not elevated by slaves - they are driven to hell. I am not sure where human users of LLMs will end up, psychologically. Is there a hell for robots, and can they take us with them? A question for a theologian, maybe.
The Future of Drugs
Coming back to the argument that information and productivity can be addictive, like substances, but perhaps more as practices than consumables, let's end with a prediction on where this takes us. To say that neoliberal societies (some but not all of the countries of the world) are addicted to work is redundant; neoliberalism knows nothing but work, so to say it is addicted to work is like saying the Christian Church is addicted to metaphysics. Neoliberalism's raison d'etre is to extend the realm of human labor into all human practices.
Technology has allowed for the processing of human practices - knowing and doing things - into forms of labor. It is work to read the news; I do it sometimes and nearly always feel tired afterwards. That my attention to articles is observed by trackers, processed by algorithms, and sold to advertisers is just one particularly clear example of how my wanting to know things can be turned into a revenue stream for someone. This has been true since newspapers, but newspapers are less fun to read than websites with streaming video, pop ups, hyperlinks, and comments sections.
The processed nature of drugs is what makes them distinctively dangerous to human bodies. Alcohol, caffeine, cocaine, and marijuana...to name a few...are all famous and common thanks to their successful progression as highly processed and consumable substances. Were they hard to consume, the market would be smaller. Look no further than needle-delivered heroin for an example: how many of us would be eager to experiment with a needle between our toes for the promise of euphoric sensation? That heroin is used at all is testament to its incredible power.
Social media and LLMs are not at all like shooting up between your toes, or into the soft of your inner elbow. Using them, in my opinion, feels more like the chirpy buzz of a hot cup of coffee, the gasp for air after drinking a sugary bubble drink, or leisurely glass of wine after dinner. There's something soothing, in other words, in consuming and producing large amounts of information daily (and for some of us, getting paid to do it). These drugs are 'low key'. They're soothing in their monotony.
The future of drugs is work. Or, the future of work is drugs.
We are working on making our jobs nihilistic. David Graeber knew it, and in death still sells books on it. Mark Fisher likely knew it, but hated it too much to take it. If our jobs are meaningless but pleasurable to do, all the better for neoliberal systems that can only imagine more work in the future.
Self-reflection
I've said too many sweeping and judgmental things not to place myself in this.
I could very easily use LLMs to expand my working capacity. Well, 'expand'. Specifically, as a lecturer, I could use LLMs to make my slideshows, my quizzes, my essays prompts, and even do my essay grading for me. Hell, I could assign an essay prompt every week or month to students (and they'd have to do it to pass), then use an LLM to grade what I was given. I could fill my course with work, more and more of it, and process the grades through an online portal, and throw those grades out at the end.
There is a real, if bemusing, temptation here. If one wanted to fill their world with work, LLMs are ready to do the job of the finer details.
We face a dilemma with LLMs that alcoholics face in grocery stores: there it is, there's the stuff, I could have just one or two, it's been a while, I'm under control, it wouldn't be too much...
Not discussed just yet, but always the subtext, is the fact that drugs are defined by their addictiveness. Not being addictive, they'd simply be substances (or in the case of knowing and doing things, experiences). Music is hard to label a 'drug', since our enjoyment of it tends to vary by so many other factors. I will listen to a new song I love 20-30 times in a week before getting tired of it...and maybe never listening to it again. It's power will wear off. A good beer, though, will get me drunk at any time.
Are my proposed drugs of information and productivity more like music, or more like alcohol? My view should be clear: in their highly processed, technologically slick method of experiencing them in the 2020s, they are more like alcohol. These are experiences laden with dopamine hits: memes and magic boxes. This is why we like social media and LLMs: they are not more useful or necessarily better than the way things were done before. They are just more pleasurable and easier.
This is their similarity to alcohol. It is easier to 'relax' through drinking than to shut off one's phone, sit or lie down, and be with one's thoughts and reflections, unwinding and unpacking the day. Some people surely genuinely relax after work without chemical assistance. Vast numbers turn to substances, though, whether that be e-cigarettes, tea, wine, sugar.
Turning work into a drug means being able to do more work. Doing more work, in a neoliberal society, is highly valued. It is perhaps the highest valued activity, above raising kids, having friends, and even being healthy. Look no further than the thousands of wealthy middle managers and empty suits who die wealthy and unhappy in middle age. These are the heroes of neoliberalism - not the freakish billionaires nor the actual-economy-workers, but the tragicomic rat racers in the middling wasteland of talent and job titles. And these are the people who will be most addicted to the new drugs, and spread them to the rest of us.