AI Could End Up Too Cheap to Control

Capital markets have priced Anthropic and OpenAI near $1 trillion each on the assumption of enormous future profits. The argument here is that the opposite may happen — that inference gets cheap enough to escape any chokepoint regulators could use.