The “good old days” fallacy assumes that the past is always better. That’s not what I mean. For every Sgt. Pepper squeezed out of a four-track, there were mountains of forgettable slop pressed on the same machines. The real point is that history isn’t a steady line of progress. A medieval town might not have been paradise, but it was probably preferable to the trenches of 1914. Some ages give you space to live, some crush you flat. The same holds true for art: you can live through renaissances and you can live through dark ages. And the dark age we’re in now isn’t defined by the year on the calendar but by the broken relationship between the artist and the tool.
We’re surrounded by abundance without architecture, possibility without price. Every tool is available, but the conditions that once gave rise to coherence and lasting work are absent. That doesn’t mean the past was better in some golden-age sense — it means that right now, we’re in one of those trench periods. And if you don’t name it as such, you can’t climb out of it.
The dark age we inhabit is structural, not temporal. It stems from a fundamental shift: tools that once required mastery now operate independently, creating a disconnect between human intention and artistic output.
We have achieved material abundance in creative resources. Every color, every sound, every word is instantly available. Digital tools can generate images indistinguishable from paintings, compose music in any style, write in any voice. Yet this abundance lacks organizing principles—the constraints and limitations that historically shaped artistic expression into coherent forms.
The traditional apprenticeship model is obsolete. A painter once spent years learning to mix colors, understanding how pigments behaved on different surfaces. A musician developed calluses from practice, intimate knowledge of their instrument’s quirks and capabilities. These physical relationships with materials created natural boundaries that forced choices, developed taste, and built technical foundation.
Current tools eliminate this friction. AI can execute any visual concept without understanding composition, light, or form. Software can produce complex musical arrangements without knowledge of harmony or rhythm. The result is technically proficient output divorced from the experiential knowledge that traditionally informed artistic decisions.
This creates a specific problem: when everything is possible, selection criteria become arbitrary. Without the natural limitations imposed by material constraints or technical skill requirements, artists lose the framework that historically guided creative decisions. The abundance becomes overwhelming rather than liberating.
The economic model has shifted accordingly. Value once derived from scarcity—the rarity of technical skill, the time investment required for mastery, the physical limitations of materials. Now the bottleneck has moved from execution to conception, from craft to network effects curation.
Contemporary creation resembles industrial management more than traditional artmaking. Artists become prompt engineers, selecting from algorithmically generated options rather than directly manipulating materials. The relationship is supervisory rather than tactile, conceptual rather than embodied.
This is lateral movement, not advancement—a reshuffling of capabilities that doesn’t constitute progress in any meaningful sense. We’ve traded one set of limitations for another, gaining computational power while losing embodied knowledge.
Dark ages are indeed technological impositions. The fall of Rome wasn’t just political collapse—it was the loss of engineering knowledge, the forgetting of concrete formulations, the abandonment of architectural principles that had taken centuries to develop. The Islamic conquests didn’t just change borders; they severed the transmission networks that connected Greek philosophy to European monasteries. Printing didn’t democratize knowledge initially—it flooded markets with forgeries and propaganda, drowning signal in noise.
Each dark age follows the same pattern: new technology disrupts established knowledge systems faster than replacement frameworks can develop. The technology itself isn’t inherently dark—it’s the gap between disruption and reconstitution that creates the darkness.
The movable type press destroyed scriptoriums where monks had spent lifetimes perfecting calligraphy, developing deep relationships with text through laborious copying. The industrial revolution eliminated craft guilds that had preserved technical knowledge for generations. Radio and television displaced oral traditions that had maintained cultural continuity for millennia.
Our current disruption follows this historical template. AI tools are dismantling the apprenticeship systems that connected artists to their materials, the editorial hierarchies that filtered signal from noise, the economic structures that supported sustained creative development. The tools work, but the surrounding architecture of knowledge transmission has collapsed.
The darkness isn’t in the tools themselves but in what we’re losing while learning to use them. Every dark age ends when new institutions emerge to organize and transmit knowledge within the altered technological landscape. But the transition period—where old systems are defunct and new ones haven’t solidified—is genuinely dark. We’re in that transition now.
The question isn’t whether we’ll emerge from it, but what knowledge we’ll have permanently lost by the time we do.
The great albums of the Architectural Age weren’t great despite their limitations. They were great because of the specific creative behaviors those limitations enforced: pre-production, performance, and commitment.
So the lesson isn’t to mourn the past. It’s to reverse-engineer its principles.
The Iron Law of System Evolution
Digital tools dominate markets through superior capital efficiency, not superior output quality. An AI system generating a thousand images daily operates with lower marginal costs than human illustrators producing individual works. This economic advantage drives adoption regardless of aesthetic merit.
Venture capital allocates resources based on scalability metrics. Systems that can serve unlimited users without proportional increases in labor costs receive funding. Traditional art production—requiring individual human time per output—cannot match these unit economics. The technology wins through financial mathematics, not creative superiority.
This process systematically displaces accumulated knowledge. Renaissance techniques developed over centuries—perspective systems, color relationships, compositional principles—become economically irrelevant when algorithms approximate these effects without understanding underlying principles. The knowledge exists but loses transmission pathways as institutions supporting traditional training become financially unsustainable.
Historical dark ages follow similar patterns. Roman engineering knowledge disappeared not because it was inferior to medieval techniques, but because the economic and institutional systems supporting its transmission collapsed. Islamic preservation of Greek texts occurred through economic incentives, not cultural sentiment. When those incentives shifted, the knowledge networks dissolved.
Current technological systems compress centuries of artistic development into training datasets, then reproduce surface patterns without underlying comprehension. This isn’t cultural vandalism—it’s resource optimization. The market rewards speed and scale over depth and understanding.
Renaissance apprenticeship models—master-to-student knowledge transfer, incremental skill building, innovation through deep technical foundation—cannot compete economically with systems that bypass the learning process entirely. The traditional model requires time investment that contemporary economic structures don’t support.
Knowledge preservation requires economic viability. When new technologies undercut the financial basis for traditional knowledge transmission, the knowledge disappears regardless of its intrinsic value. This is how dark ages function: not through catastrophic loss but through gradual economic displacement of knowledge-preserving systems.
The Middle Ages didn’t “win” over Rome—they emerged from Rome’s institutional collapse and offered more efficient resource allocation for the conditions that followed.
Roman civilization required massive infrastructure maintenance: aqueducts, road networks, urban centers, professional armies, complex administrative bureaucracies. This system worked when it could extract sufficient resources from conquered territories to fund its operations. When expansion stopped and resource extraction declined, the maintenance costs exceeded available capital.
Medieval systems operated with lower overhead. Feudalism distributed governance to local lords who managed smaller territories with direct agricultural surplus. Monasteries preserved knowledge through self-sufficient communities rather than expensive urban institutions. Castle-based defense required fewer resources than maintaining professional legions across vast frontiers.
The technological shift wasn’t qualitative improvement—it was adaptation to resource scarcity. Medieval agriculture was less productive than Roman agriculture, medieval engineering inferior to Roman engineering, medieval administration cruder than Roman bureaucracy. But medieval systems could function with less capital input.
Christianity provided ideological infrastructure that justified this resource reallocation. Instead of expensive public works projects, resources went toward spiritual salvation. Instead of costly urban amenities, populations accepted rural subsistence in exchange for eternal rewards. The religion didn’t cause the transition but legitimized the economic necessity of operating with less material abundance.
Roman knowledge systems required concentrated resources: libraries, schools, professional scholars, urban centers with sufficient population density to support specialization. When these resource concentrations became unsustainable, the knowledge networks collapsed regardless of their intellectual superiority.
Medieval knowledge preservation occurred through monasteries because monastic communities could sustain themselves through agricultural production while maintaining minimal literacy. This wasn’t better scholarship—it was economically viable scholarship under resource-constrained conditions.
The “Dark Ages” represent the gap between Roman institutional collapse and medieval institutional consolidation—the period when neither system could effectively organize knowledge transmission or resource distribution.
That’s why dark ages are cheaper for governance models (tongue in cheek). They are the ultimate austerity program.
Why maintain expensive Roman infrastructure when you can let the aqueducts crumble and tell people that suffering builds character? Why fund public education when monasteries will preserve just enough literacy to keep the tax records? Why support professional armies when local warlords will provide security in exchange for agricultural surplus?
Medieval governance operated on venture capital principles: minimal overhead, maximum extraction, externalize all costs to the peasantry. Feudalism was essentially a franchise model—grant territorial licenses to local operators, take a percentage of revenue, avoid direct operational expenses.
The Church provided perfect ideological cover for this cost-cutting. Poverty became virtue, ignorance became faith, technological regression became spiritual purity. Roman concrete formulas were lost not through accident but through calculated neglect—why invest in permanent infrastructure when temporary solutions require less upfront capital and create ongoing revenue streams through constant repair needs?
Even better, dark age governance models convinced populations to be grateful for the degraded service. Roman citizens expected clean water, paved roads, professional security. Medieval subjects thanked God for muddy paths and hoped bandits wouldn’t rob them this season.
Modern platform capitalism follows the same template: eliminate expensive human expertise, replace professional services with algorithmic approximations, convince users that degraded output is actually “democratization.” Why pay illustrators when AI generates adequate content? Why maintain editorial standards when engagement metrics optimize themselves?
Dark ages aren’t historical accidents—they’re the natural result of optimizing governance costs. The only question is whether populations will accept the service degradation in exchange for lower overhead expenses passed on as “accessibility” and “innovation.“
So probably what we can expect from LLMs is to produce at best some input like Bede, the English monk from the VII and VIII century, important in its own time but forgettable in the context of what came before it and what came after. Bede was genuinely important for his historical moment—the Ecclesiastical History preserved crucial knowledge during a period when institutional memory was fragmenting. But it’s essentially compilative work, synthesizing and transmitting existing sources rather than generating new understanding.
LLMs are performing exactly that function: sophisticated compilation and recombination of existing cultural material. They’re the Venerable Bede of the digital dark age—valuable for preservation and accessibility, but not for breakthrough insight or genuine innovation.
And like Bede, they’re products of their constrained circumstances. Medieval monks couldn’t recreate the philosophical sophistication of Augustine or the historical methodology of Tacitus, but they could maintain basic literacy and organize available sources. LLMs can’t recreate the embodied knowledge of master craftsmen or the breakthrough insights of original thinkers, but they can efficiently process and recombine vast amounts of existing material.
The comparison extends to reception history too. Bede was enormously influential in his time—practically every medieval history drew from his work. But once genuine historical scholarship resumed in the Renaissance, his limitations became obvious. He was a crucial bridge, not a destination.
We’re probably looking at a similar trajectory. LLMs will dominate content production during this transitional period because they’re adapted to current resource constraints and technological capabilities. But when new institutional frameworks develop that can support deeper knowledge transmission and original thinking, the algorithmic outputs will seem as crude as Bede’s historical methods appear to modern scholars.
The real question is how long this transition period lasts, and what percentage of accumulated human knowledge gets compressed into training data versus genuinely preserved through living practice.
Leave a Reply