You don’t need a shadowy government cabal to explain Alexa persisting too long in the strategy described in the article. Just a powerful chief executive with an attachment to the product, the sunk cost fallacy, a company with too much revenue to know what to do with, some gameable downstream impact metrics, the inertia of large institutions, and empire building.
It still could’ve become an incredible product if they’d gotten on the LLM chat/agent train early enough, but alas. I mostly use mine to set timers when my hands are full in the kitchen.