3 ms·
Google's business needs aside - aren't modern phones powerful enough to process chat content locally, without exposing anything to the outer world? Or it's stil
by drdaeman 10y ago
Google's business needs aside - aren't modern phones powerful enough to process chat content locally, without exposing anything to the outer world? Or it's still more CPU or memory-hungry task than a phone could provide? (I don't have any knowledge/experience in modern computer learning, so genuinely curious.)
- maxerickson 10y agoIt's the data. If you are in New York and chat "Where should we eat?", the phone has to either have a massive amount of data on hand or request some data to make local suggestions. Then you have to balance how current the data is and storage and so on against the elaborate privacy protection. Google knowing that some person is near some intersection probably isn't much different than Google knowing some person is in NYC. Fresh results about the intersection will naturally be faster to return than fresh results for a larger area that then have to be filtered down on the device. I guess this is a good reason to add POIs to OpenStreetMap.
- drdaeman 10y agoWell, analyzing locally and sending a query to Google Maps/OSM/whatever that basically says "food + New York" is one thing. For extra privacy, there can be even a nice opt-in that actually works with AI breaking into conversation with a "Hey, want me to suggest a good restaurant? Just tap here and I will look up some!" I just wonder if today phones are powerful enough to process texts locally and not disclose the raw unprocessed data to the third parties.
- yellow_postit 10y agoIt is likely less an issue of processing locally but a need to build a collection of data to train future versions of the service. To do that you need to see how users are trying to use the "bot", what it doesn't handle well, etc. That lets them improve the models doing the NLP.