> In the end, even if the features aren't perfect, they still raise the bar for competitors Honestly I'm not sure that they do. Everything I've seen with notif

> In the end, even if the features aren't perfect, they still raise the bar for competitors Honestly I'm not sure that they do. Everything I've seen with notification summaries for instance has given me the feelings of "wow I guess I'm really not missing much". LLMs as an answer engine has a big benefit that it feels fast and fluid even when its wrong (and many won't bother verifying) but with notification summaries most users in messaging contexts will eventually go back to the conversation and see the responses in full detail. Mistakes in that context are identifiable by mainstream users as having made the product worse.