Six Dead Men on AISix Dead Men on AIMachiavelliMachiavelliOpenAI Is a Textbook forOpenAI Is a Textbook for

OpenAI Is a Textbook for *The Prince*OpenAI Is a Textbook for *The Prince*

Here is where his lens cuts deepest, because he wrote the case study five hundred years early.

Chapters 6 through 9 of The Prince are about the new prince — the man who acquires power not by inheritance but by his own ambition, and who therefore faces the hardest problem in politics: consolidating control over an apparatus that does not yet recognize his authority. His central warning is about founders who rely on the goodwill of those who elevated them. "He who becomes prince through the favor of the people must keep them friendly, which is easy... but he who becomes prince by the favor of the great" must master them before they master him.

Now look at OpenAI. Founded in 2015 as a nonprofit, with a charter, a capped-profit hybrid bolted on in 2019, a board ostensibly empowered to shut the whole thing down for the good of humanity. In November 2023 that board fired Sam Altman. Within five days, ninety-plus percent of staff threatened to walk to Microsoft, and the board that held the formal kill switch was itself killed. Machiavelli would recognize this instantly. The new prince had been temporarily deposed by the institution that created him — and he won not through the charter, not through governance, but through virtù: the seizing of fortune's moment, the loyalty of his own arms (the employees), and the decisive counter-move. The man returned stronger; the constraining institution was gutted and reconstituted in his image. The subsequent grinding conversion from nonprofit toward a conventional for-profit is simply the new prince doing what The Prince says he must: dismantling the structures of the old regime that could ever again limit him.

The "safety" framing misreads this entire episode as a governance failure. Machiavelli would call it a governance success — power found its level. The lesson he would draw for policymakers is brutal: never build your AI oversight regime on a nonprofit charter or a voluntary board, because those are fortunes, not arms. They evaporate the moment a determined prince decides to test them.

Here is where his lens cuts deepest, because he wrote the case study five hundred years early.

Chapters 6 through 9 of The Prince are about the new prince — the man who acquires power not by inheritance but by his own ambition, and who therefore faces the hardest problem in politics: consolidating control over an apparatus that does not yet recognize his authority. His central warning is about founders who rely on the goodwill of those who elevated them. "He who becomes prince through the favor of the people must keep them friendly, which is easy... but he who becomes prince by the favor of the great" must master them before they master him.

Now look at OpenAI. Founded in 2015 as a nonprofit, with a charter, a capped-profit hybrid bolted on in 2019, a board ostensibly empowered to shut the whole thing down for the good of humanity. In November 2023 that board fired Sam Altman. Within five days, ninety-plus percent of staff threatened to walk to Microsoft, and the board that held the formal kill switch was itself killed. Machiavelli would recognize this instantly. The new prince had been temporarily deposed by the institution that created him — and he won not through the charter, not through governance, but through virtù: the seizing of fortune's moment, the loyalty of his own arms (the employees), and the decisive counter-move. The man returned stronger; the constraining institution was gutted and reconstituted in his image. The subsequent grinding conversion from nonprofit toward a conventional for-profit is simply the new prince doing what The Prince says he must: dismantling the structures of the old regime that could ever again limit him.

The "safety" framing misreads this entire episode as a governance failure. Machiavelli would call it a governance success — power found its level. The lesson he would draw for policymakers is brutal: never build your AI oversight regime on a nonprofit charter or a voluntary board, because those are fortunes, not arms. They evaporate the moment a determined prince decides to test them.