This also plays out over time and assuming future LLMs will be trained ever more on the generative output of todays LLMs we could end up communicating using a L

This also plays out over time and assuming future LLMs will be trained ever more on the generative output of todays LLMs we could end up communicating using a Language itself that will become frozen in time as it cycles through the same feedback loop. As John Culkin said: “We shape our tools, and thereafter they shape us.”