The Slop That Lost

The AI slop did not.

This goes in the very long file labeled "confusing broadcast for coalition."

Pratt's campaign had a genuine origin: he lost his home in the 2025 Pacific Palisades wildfire, one of the disasters that defined Karen Bass's first term. The anger read as real even when the content was obviously machine-generated. That combination — authentic grievance, enormous social reach, cheap production — looked, briefly, like a new political formula. More than a million X followers, viral content, no political infrastructure. The question was whether viral reach could substitute for ground game.

It cannot. This has been tested repeatedly and the answer is the same.

The structural problem with AI-generated content as a political tool isn't aesthetics, though the aesthetics are often bad. It's that virality is a broadcast metric, not a mobilization metric. A video with five million views has been seen. It has not been acted on. The people who watched Batman Pratt fight Joker Bass were distributed across the entire platform — global, demographically incoherent, connected to nothing in Los Angeles's precinct-level political geography. They were an audience. Audiences don't vote as a bloc; constituencies do.

Ground games exist because reaching someone once is not the same as moving them. The Los Angeles Democratic machine has had decades to build the kind of infrastructure that turns attention into turnout. Pratt had months of viral content and no equivalent.

This is the latest version of a pattern that runs through every new broadcast medium. Radio demagogues in the 1930s discovered they could build mass listenership without converting it to durable political power. Early television produced insurgent candidates who mastered the format and lost to operators with ward bosses and unions. Social media in 2016 appeared to break the equation — it hadn't; it produced one unusual election cycle before the underlying mechanics reasserted. AI-generated content is faster, cheaper, and more scalable than anything before it. This makes it better at reach. It makes it no better at organizing. The gap between those two things is where campaigns die.

The story being told about Pratt's loss — that AI slop failed, that voters rejected the format — is the wrong frame. AI slop performed exactly as broadcast media always has: captured attention at scale and failed to convert that attention into coordinated local action. That's not a failure of the technology. It's the technology working as designed, being mistaken for something it isn't.

The question isn't whether AI political content will keep spreading. It will. The question is whether anyone drawing lessons from the Pratt campaign understands the difference between making something go viral and building something that holds.

Coalitions are load-bearing. Audiences are not.