Seeing Like an AI Company

OpenAI's industrial policy proposal is a legible artifact: it wants the state to see the AI economy the way the company already does. So I handed it to six dead scholars. Scott on legibility, and five others who each name a problem the document works hard not to name. The interesting part is that the six problems connect: read together they describe one project, not six complaints.

Read OpenAI's industrial policy proposal closely and the gaps aren't accidental. They're the shape of the argument. Six dead scholars, six problems the document refuses to name, all pointing at the same thing.

The piece: https://www.jakelawrence.xyz/research/seeing-like-an-ai-company

OpenAI's industrial policy proposal wants the state to see the AI economy the way the company already sees it. Neat, ordered, fundable. I gave the document to six dead scholars, starting with James C. Scott and his idea that states simplify the world until it fits their forms. Each finds a problem the text avoids. Read together, the six problems aren't a list of complaints. They describe a single project.

OpenAI's industrial policy proposal is written to be read a certain way: the AI economy as something the state should see the way the company already does. Legible, ordered, ready for subsidy. So I ran it past six dead scholars. James C. Scott on how states flatten what they can't measure. Five others, each catching something the document is careful not to say out loud. The surprise wasn't any single objection. It was that the six problems interlock. Pull one and the others move. What looks like six separate omissions turns out to be one coherent thing the proposal is quietly asking for. https://www.jakelawrence.xyz/research/seeing-like-an-ai-company

OpenAI's industrial policy proposal wants the state to see the AI economy the way the company already does. I gave it to six dead scholars. Each names a problem the text avoids, and the six problems turn out to be one.

https://www.jakelawrence.xyz/research/seeing-like-an-ai-company

Six dead scholars read OpenAI's industrial policy proposal. Each finds a problem it refuses to name. The twist: the problems connect, so you're not looking at six omissions but one project the document won't describe plainly.

https://www.jakelawrence.xyz/research/seeing-like-an-ai-company