I remember testing a protocol called DDE back in the Windows 3.1 days, and we couldn't figure out what Excel behaved so much faster than Lotus 123. Then we dug into the code and found that Windows 3.1 would free up all available memory as it launched Excel. A sneaky trick which I guess meant that users would find Excel much more responsive that 123 and ultimately help Microsoft get Office domination that it has today.
I honestly don't know why Apple can't do their own ChatGPT. They have the money / resources to then implement something like the ChatGPT 'plugin' architecture so if someone wants to turn on their lights or get specific 'private' information that apple cannot give as 'training' data. i.e. callout to 'private' API's if appropriate to a sensible LLM generated response
I just load my audiobooks collection onto nextcloud and when I want to listen to an audiobook I share the file with PocketCasts which has full m4b support with chapters and covers and has excellent audio support to speed up the playback and skip silence (can get through a 8 hour audiobook in about 5-6 hours