Poetry in the age of AGI
A neo Humanist manifesto
A friend asked me to share my thoughts about a post-AGI (artificial general intelligence) world.
TL;DR: What will save us in a post-AGI world is poetry.
Or more generally: in a world where machines solve problems much better than us, there will be little point in giving people a STEM-focused education that teaches them solve problems. Instead, what we should focus on is giving people an education that teaches them how to spend time in a pleasant and meaningful way, pursuing beauty while machines pursue effectiveness; while cultivating their critical sense and their sense of ownership for their future.
The challenges of AGI
But let’s take a step back.
If the scaling laws hold (and that may be a big if), at some point we’ll hit AGI, and immediately after, as AIs are put to work to improve AI, we are likely to cross into superintelligence.
Superintelligence will by definition surpass humans in all forms of problem-solving. It will be able to solve any scientific or technological problem in a fraction of the time it takes humans, and it will basically keep improving itself indefinitely. It will, of course, substantially impact our jobs. It will first eliminate the menial aspects of problem-solving activities, then the creative parts, until basically nothing remains. This applies to intellectual work — from lawyers and economists to engineers, designers, and programmers. With superintelligent systems capable of self-improvement, we may soon see humanoid robots handling physical world interactions, from plumbing to massage therapy.
The emergence of AGI poses three key challenges:
The Economic Challenge
The primary economic challenge is maintaining prosperous lives for people in a world where most will be unemployed because machines will be doing their work. AGI is likely to exacerbate extreme inequality, as those who own the compute charge companies and government to basically run anything. A solution to the economic challenge has been talked about for a relatively long time, and that’s UBI — universal basic income. People like less to talk about how we will pay for that, but that is also relatively straightforward: redistribution. The wealth produced by these machines, owned by companies and individuals, will need to be redistributed through taxation and other mechanisms to support the broader socioeconomic network. This is of course likely to work only with some measure of international coordination. Perhaps the types of GPUs on which AGI will run could be subjected to an ad-hoc taxation scheme, whose revenues can go to offset its environmental impact and the social displacement it will create.
The Political Challenge
The political challenge is already emerging, even before true AGI arrives. The cost of creating and sharing content is being reduced to near zero. How do we preserve our democracies and political institutions in a world where truth becomes increasingly difficult to discern? With deepfakes and AI-generated content, any narrative can be propagated indefinitely by bots. It is virtually guaranteed that bad actors will try to exploit this mechanism to influence voters. Before our eyes, an increasingly polarized public opinion is starting to diverge on very basic facts. If we keep the current attention-monetizing mechanisms in place and fill them with AI generated content that is increasingly able to hack our attention, the consequences are unpredictable but likely grim.
The psychological and social Challenge
The third challenge concerns meaning: In a society where machines handle most work and, let’s assume as a best case scenario, people live off UBI, what will people do with their time? How do people find purpose in their lives? We’re already seeing AI taking away many enjoyable aspects of people’s jobs. How do we make them feel useful, loved, cherished as this process gains pace? Here too, if the present is any guide, companies will likely compete to capture people’s attention, coming in to fill our days with an endless series of quick-fading dopamine boosts. This is likely to narcotize is into contentment for a while, but the effects of alienation and depression will quickly emerge. People might fall prey to various forms of addiction — from alcohol and drugs to pornography, gaming, and gambling. Eventually, unrest and crime might spike. This all sounds very gloomy, but it is also quite possible that AGI might improve the life for many of us, as we will become increasingly able to create virtual worlds where life is perfect. Of course, even in a future where anyone can plug into virtual worlds that provide endless pleasure and stimulation for the long run, the possibility that one day we might take off our VR set and find a world that has become a totalitarian dystopia while we were anesthetized is a real worry.
The unlikely comeback of humanities
Even if there was the political will to stop AI’s progress, all incentives are in place for its continuation at an ever increasing pace. Unless we hit some hard limit dictated by the current technology, data, energy and compute availability, AGI is likely to eventually come. So how could humanity prepare for the groundswell that it will inevitably unleash?
The answer, of course, is that we should leverage what is uniquely human to find meaning and shape a world that still makes some kind of sense.
We have seen that the economic challenge might be sorted out with some form of redistribution and guaranteed income. This is a titanic challenge on its own of course, but it’s still the easiest part of the problem. Once that is overcome, humanity will be facing something completely new: once rid of the shackles of the need to work, our society will be shaped by how we choose to occupy our time. What content will we consume? How will we interact? How will we make decisions?
I propose that a solution to the political and psycho-social challenges might come from an unexpected place: humanist education, coupled with regulation that levels the attentional playing field between Russian novels and cat videos.
From problem solvers to meaning makers
What are the humanities about? I think teaching the humanities — literature, the arts, music, is teaching people to recognize, enjoy and craft beauty in ever more sophisticated ways.
Machines will surely be able to create beautiful things. They can already write half decent stories and create pretty impressive pictures. So what is uniquely human about humanities?
First, humanities are about recognition, appreciation, enjoyment first. This remains uniquely human: short of becoming conscious, AI may never appreciate the qualia of beauty. More than that,
in a post-AGI world, humans will effectively shift from primarily being busy solving problems to primarily being focused on experiencing beauty.
Skillful making will still be held as valuable, but relishing experiences will be an even more central source of meaning.
Additionally, what happens in a world where AI can create content that is as good, and somewhat possibly more “perfect” than humans? What happens is that machine-made content quickly becomes 99% of all content out there. For non-utilitarian, artful creations, we’ll be quickly fed up with this perfection. We’ll see through the canvas, losing patience with machine-generated art, no matter how beautiful it might be. We’ll seek out human-made content because it’s human made: in a sea of sameness perfection, what we will choose to experience is art with imperfect, but human will behind it.
Tolstoy vs Mr. Beast
One might argue that, if life will be about experiencing, any entertainment will do. I don’t think so. We all know from experience that some kinds of entertainment are short lived and leave us nauseous, whereas other kinds leave us feeling enriched. It is hard to point to why this is the case, but I think it has to do with the fact that, while a cat video or a dull video game only speaks about itself, an great novel (or an exceptional video game, of course) tells us something about life in general, something we can then appreciate once we are taken out of that experience.
The humanities are about learning how to appreciate works that are deemed exceptional, and through this knowledge, find beauty wherever it is to be found. Adequate knowledge allows us to perceive beauty in increasingly sophisticated ways: to the untrained ear, a complex song or symphony might sound like a jumble of notes; to the expert, a masterpiece. The humanities are really the obvious case here because their entire focus is on beauty, but of course, we can still train biologists to see the elegance in organisms, botanists to relish in the subtle distinctions between different plants, mathematicians to delight in sets and symmetries. The hard sciences can be an immense source of beauty, but we’ll be doing them in a radically different way: once the problem solving aspect is gone, the scientists will look at lab results like the critic looks at a painting.
One might think this is an elitist perspective. It isn’t. While people could spend their existence playing addictive games or consuming entertainment, such pleasures often lead to a spiral of meaninglessness. History shows us that people can derive extremely meaningful lives without solving practical problems, simply by delving deep into artistic and cultural pursuits.
At the same time, the humanities can help reinforce pro-social values. Literature, art, and philosophy have historically served as vehicles for developing empathy, critical thinking, and moral reasoning. When we engage with great works, we’re not just appreciating their aesthetic value — we’re participating in a centuries-old dialogue about what it means to be human. This engagement helps develop the emotional intelligence and ethical framework needed to navigate complex social dynamics. In an age where AI might excel at optimization and efficiency, these deeply human capacities for moral reasoning and empathetic understanding become even more crucial. They form the bedrock of social cohesion and help us resist the atomizing forces of technological isolation.
Protecting our attention
But how will people be willing to engage with deeply enriching content when a Mr. Beast video dispenses dopamine far more generously?
Like any vital, limited resource our attention should be protected. We are used to think of attention as a free market. At some point, the markets figured out that attention can be assigned monetary value, so it became a commodity. Companies are constantly competing for our attention, and there are basically no rules to what they can do to get it or keep it. This has given rise to a generation of applications that are designed from the ground up to keep our attention on them. Fear of missing out, variable rewards, endowment effect, social pressure: a whole panoply of psychological tools is deployed to keep our eyeballs on screens that serve us ads that someone pays for. In an age where AI can further perfect these addictive mechanisms, not to mention generate addictive content at zero cost, the situation can only get worse. I could mention that app addiction and heavy smartphone usage in general are correlated to a whole host of mental diseases, but there is no need to bring disease into the picture. We all know how it feels to see in hindsight half an hour of our life down the social media drain: the feeling of waste, guilt, restlessness, meaninglessness. We need to do everything we can to make sure our post AGI future is not reduced to a few flashes of awareness in between hours of mindless scrolling. We cannot help ourselves, so regulation must come into play.
Just like toxic foods are prohibited and unhealthy foods are labelled as such in many countries, so we should do with our apps. It’s not a matter of censoring content, god forbid. It’s a matter of forbidding the addictive mechanics, just like India has done for dark UX patterns. We cannot let the humanities compete in a free market with cat videos and porn, because that would be unfair competition. So just like we ban dumping or cocaine, we should find mechanisms to limit the extent to which people might want to consume addictive content.
The advent of AGI will make it possible to spend our lives living memorable, enriching, deeply satisfying experiences of awe, gratefulness, love, connection. But this will only be possible if we’ll find a way to protect our attention — which is the thing our life ultimately consist of — from the ever increasing threats of slop, mindlesness and addiction. We can hope for some regulation on the matter, but in the meantime, keeping your phone in your pocket a bit more won’t hurt.
