Hacker News new | past | comments | ask | show | jobs | submit login

Serious question: is it feasible to implement a kind of loose voice 'fingerprint' to prevent this kind of thing? Will/could Alexa know who's talking to it?



I really want an 'Alexa, stop listenting' command. There's a button on the top that mutes the mic and puts a red ring around them, but when I have people over, it's not a great environment to use voice commands anyways.

'Everyone be quite so I can shout across the room to change my music'


A workaround would be to mute the device itself, and then use the remote (which has its own mic, and works well in noisy environments since you just hold it closer to your mouth).


but then I just have to carry a remote while I'm having a party.


It's not a bad idea. If you're hosting, it lets you change the music without interrupting your guests.


Yes, definitely. It adds complexity in that now you have another source of both Type I and Type II errors (failing to wake up, waking when it shouldn't). Voice ID itself is far from settled science to do well, so it would be a tradeoff.


Theoretically yes you can fingerprint voices. The questions are:

1) Can you do this on the Alexa servers efficiently

2) Do you want to? Seems like setting it up could be a hassle. Right now there is zero friction and it just works.


It's not necessary anyways, as you have to provide a pin to actually order things.


Short answer: yes.


Long answer? i.e. is it possible out of the box or is it possible in principle, if anyone actually builds it?




Consider applying for YC's first-ever Fall batch! Applications are open till Aug 27.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: