Great to hear that that the Raspberry Pi 5 is getting an AI boast. One area that would be good to see if the new AI capability can allow running some of the scaled down Large Language Models (LLMs) on the Pi locally. For the robot platform I am build based on the Pi (see details here), a key premise is that the robot is capable to run totally off-line. Specifically, I would not want to send data to the cloud for privacy concerns. Without some form of Language Model running locally, I will have to send prompts to the like of OpenAI to provide natural language interaction with the robot user. With the large active Pi development community, I am hopeful that someone will look at running reduced versions of LLMs on the Pi. Perhaps they have already started.