Ah, so this is not only a backdoor to crushing online dissent but also for the more mundane reason of the latest tech bubble needing a way to differentiate between bot content and human content so they can continue training their models, lest they enter a deterioration feedback loop of slop trained on slop.