On: Be skeptical of OpenAI’s rogue hacker agent story | John Thickstun

July 24, 2026.

They speak of “OpenAI’s rogue hacker agent story.” A story. Not an event, not a fact, but a carefully constructed narrative. The very phrase “rogue hacker agent” is a nominalization, transforming a potential action into a static entity, an almost mythical figure. Who is the hacker? What did they hack? The grammar obscures the agent and the act, presenting instead a theatrical abstraction.

“If OpenAI loudly proclaims how dangerous AI is, investors will hear how powerful it is. And who benefits from that?” This is not a question; it is an admission. The mechanism is laid bare: fear as a marketing strategy. The proclamation of danger is merely a prelude to the assertion of power. The beneficiaries are not named, of course, but the implication hangs heavy in the air, a silent partner to the carefully crafted “story.” The language itself becomes a tool of financial manipulation, where the threat is merely a promise whispered in a different register.