The panic over AI-generated content feels urgent and unprecedented, but it's actually following a well-worn historical pattern. According to a16z crypto's latest analysis, every time technology dramatically lowers the cost of producing media, the same cycle repeats: a flood of cheap, derivative work arrives, critics declare culture has hit rock bottom, and defenders of "good taste" emerge to draw the line between quality and garbage.

What Does History Tell Us About AI and "Slop"?

The comparison starts in 18th-century London, where Grub Street earned its notorious reputation as a haven for hack writers and failed scribblers willing to churn out doggerel, pamphlets, and periodicals for a shilling or two. Samuel Boyse, a drunken Irish poet, became one of the era's most prolific practitioners of this low-cost content production, even pawning his own shirt to buy paper for his craft. The moral judgment was swift and harsh: respectable society equated writing for pamphleteers to prostitution.

Yet here's the catch: plenty of respectable people participated in Grub Street's economy while publicly decrying it. This pattern has repeated itself across centuries. After Grub Street came the penny press, then pulp fiction, television (dismissed as a "vast wasteland"), blogging, social media, and now artificial intelligence. Each time, production costs plummeted, the masses jumped in, and critics declared civilization in decline.

Can AI Actually Develop "Taste," or Is That Something Else Entirely?

The real intellectual puzzle isn't whether AI generates low-quality content. It's what we actually mean when we talk about "taste" in the first place. If taste consists of stable, discoverable rules, then modern AI systems trained on enough examples should be able to parrot it. A model trained on Bauhaus design principles could infer what makes something feel Bauhaus. A literary agent's AI could approximate that agent's editorial preferences. Pattern recognition is precisely what large language models excel at.

But if taste isn't reducible to stable criteria, then the word has been used far too casually. Reducing taste to personal preference ("I like this; I don't like that") is too subjective to explain why certain people repeatedly identify promising ideas, products, or creators before everyone else does. This forces a deeper question: what are we really talking about when we say someone has taste ?

a16z crypto draws on social science research to disaggregate what "taste" actually comprises:

• Prediction: The ability to forecast which products or ideas will succeed before others do.
• Judgment: Applying consistent aesthetic or quality criteria to evaluate work.
• Social Position: Occupying a structural advantage that gives unique visibility into how ideas connect.
• Experience: Accumulated knowledge from exposure to many examples over time.
• Luck: Being in the right place at the right time when conditions align.

What Does Social Science Reveal About How Success Actually Works?

In the early 2000s, Columbia University sociologists Duncan Watts and Matt Salganik conducted a landmark experiment that challenges the idea that taste predicts success. They created an artificial online music lab where thousands of participants listened to and downloaded songs by unknown bands. The researchers divided listeners into eight separate groups, each with different information about how popular each song was.

The results were striking. When social information was hidden, downloads were evenly distributed across songs. But when people could see what others were choosing, outcomes diverged dramatically. A song that became a runaway hit in one experimental condition performed poorly in another, even though the underlying quality was identical. This reveals a fundamental truth: once quality clears a basic threshold, success is heavily shaped by social influence and path dependence, not by quality alone.

"The success of Harry Potter was a fluke," Salganik observed, noting that even informed experts cannot reliably predict breakout successes among strong candidates.

Matt Salganik, Columbia University sociologist

This finding demolishes the idea that taste means predicting success. Experts in a domain, like television executives, are reasonably good at filtering out obvious failures before they happen. But they cannot reliably predict which strong candidates will rise to the very top. Success is not necessarily dictated by quality alone.

How Does "Brokerage" Explain Innovation and Taste?

Another body of research offers a different lens on what taste actually is. Sociologist Ron Burt spent decades mapping social networks inside organizations, asking not whether individuals were brilliant, but how their positions within networks affected their performance. He compared communication patterns with measures like promotion, compensation, and managerial evaluations.

Burt's finding was counterintuitive: the most innovative employees were rarely embedded in a single, tightly knit community. Instead, innovators occupied what he called "structural holes," the gaps between otherwise disconnected groups. Because they moved in different circles, these "brokers" encountered information that others didn't and, more importantly, saw problems from multiple perspectives. Their position allowed them to recombine ideas that no single network could have produced on its own.

This suggests that taste may not be an aesthetic faculty at all, but rather a structural advantage. Innovation isn't a matter of exceptional intelligence; it's a matter of extraordinary position. Someone with access to multiple communities can see connections and possibilities that insiders within a single group cannot.

How to Think About AI, Taste, and the Future of Content

• Separate the Mechanisms: Stop treating "taste" as a single phenomenon. Instead, identify which specific mechanisms you're actually talking about: prediction, judgment, social position, experience, or luck. This clarity matters because AI may replicate some mechanisms while leaving others untouched.
• Recognize the Historical Pattern: Every technology that lowers production costs triggers moral panic and claims that culture is declining. This happened with the penny press, pulp fiction, television, and social media. Understanding this pattern helps distinguish genuine concerns from cyclical anxiety.
• Focus on What Remains Scarce: Once production becomes cheap, the question shifts from whether machines can generate content to which human capacities remain scarce. If taste is structural position rather than aesthetic judgment, then the scarcity lies in access to diverse networks and the ability to broker ideas across them, not in the ability to recognize quality.

The a16z crypto analysis suggests that the real value in a world of abundant AI-generated content won't come from machines that can "taste" better, but from humans and systems that can occupy structural positions connecting otherwise disconnected domains. The panic over AI slop may be misplaced; the real competitive advantage lies in brokerage, not in taste.