Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[HA Dave]

..... must be pulling in enormous amounts of data just from domestic communications. I don't see how they can store it, categorize it, catalog it, decide what to delete, etc. etc. etc.

One word:  metadata

Just like if you check you Alex app... it has (in a text form) all your recent commands/requests of your echo device. Or... if you speak your texts on your phone... you words are converted to text.

Using proper software.... metadata can also contain symbolic text that would reference accents, cadence of speech, volume, stress levels, keywords, ETC.. Each encounter would (just like a simple HTML) only require maybe 60KB (yes KB). Not much of a problem to store. And I am sure there are algorithms that could compress that as well. 

The organization of such text/data would be done in the exact same way Amazon separates your data from my data, your shopping preferences from my shopping preferences.... credit card numbers, addresses, movie and book ownership ETC..

 

[Tuicemen]

Actually Alexa app contains all your voice commands (and what it thought were commands) as sound files too, unless you go into the history and delete them.

But then are they really deleted!

[dave w]

One word:  metadata
Using proper software.... metadata can also contain symbolic text that would reference accents, cadence of speech, volume, stress levels, keywords, ETC.. Each encounter would (just like a simple HTML) only require maybe 60KB (yes KB).

Yeah, I understand that, but the meta data is in addition to the actual audio data. It still must be huge amounts of digital data to store. I don't know if I am awed or scared. 

[HA Dave]

Actually Alexa app contains all your voice commands ....

Our/the Amazon and Google voice devices... as well as the apps and services offered with the various mobile phones.... are the tip-of-the-iceberg as to where AI will be taking us. Not to mention how others will exploit artificial intelligence.... I am sure in ways we haven't even imagined yet. IMHO.

[HA Dave]

One word:  metadata
Using proper software.... metadata can also contain symbolic text that would reference accents, cadence of speech, volume, stress levels, keywords, ETC.. Each encounter would (just like a simple HTML) only require maybe 60KB (yes KB).

Yeah, I understand that, but the meta data is in addition to the actual audio data. It still must be huge amounts of digital data to store. I don't know if I am awed or scared. 

First there is no need to keep the actual audio past the analytical stage... once it's converted to data with the metadata. This is NOT evidence... this is never going to be taken to court. I know it almost sounds like Science Fiction... but it is all conventional technology... today. 

Without compression... if every American created a daily (metadata) file page (60KB)... I think that would be around 17,000 gigabytes a day. Using specialized algorithms and such... working mostly through the slower night hours to analyze, organize, reduce,  and file. I think that would be easily manageable with a budget of a few billion.
   
Theoretically speaking.... of course.