Doctorow argues DOGE is 'vibe governance' by AI chatbot

Cory Doctorow's latest Pluralistic essay argues that the appeal of AI to rulers is not efficiency but solipsism. Leaders who treat other people as statistical artifacts find in a chatbot the fantasy of a state that executes their preferences without the friction of humans who hold their own.

The essay reads DOGE through this lens. Doctorow describes it as a project in which "a sufficiently deserving leader can be trusted to vibe-code the entire apparatus of state, checked only by his own sense of rightness," and extends the frame to Canada, calling Mark Carney a "dead-eyed technocrat" pursuing the replacement of civil servants with chatbots. He offers no agency-level detail on where DOGE has actually swapped a bureaucrat for a model; the piece is synthesis, not reporting.

The load-bearing citation is political scientist Henry Farrell and statistician Cosma Rohilla Shalizi's AI as Social Technology, published by Columbia's Knight First Amendment Institute. Doctorow draws from their argument that bureaucracy's much-maligned variability is what allows policy to survive contact with the ground. An LLM asked to execute literal instructions cannot make that adaptation. "Those who build or reform bureaucracies need to satisfice across multiple intersecting needs," Farrell and Shalizi write.

Underneath sits Ben Recht's line, which Doctorow leans on: "you can't optimize a trade-off." A single objective function cannot honor conflicting priorities from different constituencies at once, which is what negotiation inside a bureaucracy does.

Doctorow closes on the failure of Yuval Harari's 2018 prediction that LLMs would end "authoritarian blindness" by giving dictators accurate ground truth. In practice, he writes, people living under surveillance "switched all the energy they used to use to put on a good show for the secret police into putting a good show on for the chatbots." Two experts in our Who's Who directory shared the piece within a day.