Because it's faster and works with spotty/no reception?
(I'm just guessing because I can't find details on this stuff by wading through the feed.)
Also, just because it's doing local inference doesn't mean it won't be uploading data back to Google periodically. As GP said, they have a strong incentive to do this, so it's reasonable to just assume they will do this until they can prove otherwise.
And doing the inference locally, and just uploading transcripts for the most part is way more cost effective. They don't have to pay for the power on these endpoints.
Well Apple is exerting market pressure, there's been a lot of privacy backlash lately, and they can make customers feel safe while still exfiltrating their data via other services that you'll be encouraged to use by the "secure platform".
Because they calculated that taking a small L here would turn into a bigger W later. Comparatively, Apple is just trying to sell more units, not prop up its Ads business down the road with higher user counts giving more impressions.
(I'm just guessing because I can't find details on this stuff by wading through the feed.)
Also, just because it's doing local inference doesn't mean it won't be uploading data back to Google periodically. As GP said, they have a strong incentive to do this, so it's reasonable to just assume they will do this until they can prove otherwise.