Google Home - Loving it!

Status
Not open for further replies.
Originally Posted by JeepWJ19
Originally Posted by 4WD
One thing to mine data to help others sell products … quite another to manipulate mushy minds into your way of thinking when they don't even understand it's happening …
(and expect me to pay for it) …


Can you give an example of manipulating mushy minds?

Edit: just so I'm on the same page as to what you're referencing.


Ask and you shall receive! I wonder how many mushy minds will understand this, sadly, only about 10% of the worlds population (or less) ...
I wonder how many people will even bother to watch it... :eek:) Hand in there, for those with short attention spans, after 3 minutes of the speech, you will start to understand.
For the 10% who will watch the whole thing, already know it ...

Tim Cook Speech

Simple link to the speech
 
Last edited:
I for one will not have anything like this and especially from Google. They are using your info gleaned from your activities. They are attempting to quietly shift your focus and leanings. How? By using negative ads and information that makes you question your beliefs.

I'm trying not to get political here but that is one arena they focus on. Their technology is insidious.


The above linked speech with Tim Cook is well worth watching.
 
Don't all of these systems send your voice to their servers when you ask the device to do something, so they can do the voice recognition on those servers instead of the low-powered local device (and so they can spy on you)?
 
Originally Posted by emg
Don't all of these systems send your voice to their servers when you ask the device to do something, so they can do the voice recognition on those servers instead of the low-powered local device (and so they can spy on you)?

Yes. It's all cloud based.

But the underlying assumption is that they don't just transmit your voice to the cloud 24x7, but only snippets after you've spoken the "wake" phrase. So to say that "they spy on you" is inaccurate, IMO.
 
Well, 1984 is coming...

I for one want some privacy in my house, which is pretty much the last place standing today where I don't have to worry about being recorded. And while I don't expect privacy anywhere else besides restrooms, inviting these systems into my house, for a little bit of convenience they provide, is not for me.

Also, I don't trust any promises a corporation makes about collecting voice or video data. There is no way for us, or even courts to find that out. They can move the data to a center in another country. That way they can claim they don't keep it, and what they keep in another country is not relevant to the one you live in.
For all we know, the data centers in US, keep stuff from EU and vice versa.

One thing I know for sure is that these huge data centers don't come cheap, so they would not have so many of them for nothing.
 
Originally Posted by KrisZ
Also, I don't trust any promises a corporation makes about collecting voice or video data. There is no way for us, or even courts to find that out.
You can easily tell the device is not transmitting any data back to the cloud unless you speak the "wake" phrase, just by looking at your router activity.
 
Originally Posted by Quattro Pete
But the underlying assumption is that they don't just transmit your voice to the cloud 24x7, but only snippets after you've spoken the "wake" phrase. So to say that "they spy on you" is inaccurate, IMO.

Even taking Google's version of the Magic Eight Ball at its most innocuous, it still provides them with even better identification from a marketing perspective than any of their other "services." What are kids asking for. What are men asking for? Women? Even if it's only one person using the device, it's as useful to Google as requests from a static IP. They get some great marketing data out of it.
 
Consumer data? Pff no one cares about that, it's of no value. What can they possibly do with such useless information?
 
Originally Posted by Quattro Pete
Originally Posted by KrisZ
Also, I don't trust any promises a corporation makes about collecting voice or video data. There is no way for us, or even courts to find that out.
You can easily tell the device is not transmitting any data back to the cloud unless you speak the "wake" phrase, just by looking at your router activity.



Yeah, I bet they haven't thought at all about data transmission patterns and throughput not knowing anyone would try and correlate the two for this exact concern, and left it all up to chance. But seriously, they could be gruoped into the "obvious" transmissions when using it, and deciphered text strings from internal speech recognition software wouldn't appear as considerable data uploading (you don't think they'd regularly upload raw audio, do you?). Recognized speech-as-text could easily provide google with trending topics discussed around the machine, as well as discussions surrounding literally any trigger word/phrase du jour.

It's all about data gathering. Spying, if you will. The three, 75 megapixel image sensors on your standard phone aren't there to make your memories brighter and make your photography better. Those cutesy little camera "filters" that know exactly what part of your face to put the dumb little animations aren't there to help you take cuter selfies.....
 
I don't remember where I read it, but the packets of data are quite small and can be sent during regular device updates or when the device is woken up by the magic word. It's not like they have to record the entire conversations and send them over in an mp3 format. Remember that the device is always listening for that magic word that would "wake" it up. It means Google can do whatever they want with this functionality. They in fact can, if they choose to, record and send entire conversations. They can then bury the data however they like and encrypt it. All under the guise of providing better services. Once they have that data, it is up to them if they strip it from personally identifying info or not.

And I would like to think this is still pretty much in infancy stage, more is sure to come.
 
Originally Posted by PeterPolyol
Yeah, I bet they haven't thought at all about data transmission patterns and throughput not knowing anyone would try and correlate the two for this exact concern, and left it all up to chance. But seriously, they could be gruoped into the "obvious" transmissions when using it, and deciphered text strings from internal speech recognition software wouldn't appear as considerable data uploading (you don't think they'd regularly upload raw audio, do you?).
The conversion from speech to text happens in the cloud, not on the device itself, so yes, compressed audio is transmitted to the cloud.

Even if they compressed it to very low quality 4 kbps audio stream, over the course of 24 hours (if they're in fact listening and recording 24x7), this would generate at least 42 MB of data a day, or 1.2 GB per month. Looking at my router logs for the past month, I don't see anywhere close to this being used by the Google device.

However, it's not at all unfathomable that there are a lot more "wake" words than just "OK Google" that Google is not telling us about, which result in some additional audio being sent "home" but not 24x7.

And yes, I'm sure over time the local devices will become powerful enough to perform speech to text locally.
 
When I went through my PFSense logs it showed about 150kb for the week we were away in Ohio when no one was using it. From what I can gather is was looking for Firmware updates because it was only a couple of the units doing the transmission and not all of them I have.
 
Originally Posted by Quattro Pete
emg said:
So to say that "they spy on you" is inaccurate, IMO.


Google's entire business model is based on spying on us. Amazon and Apple, less so, but Amazon still wants to know what things you might want to buy so they can recommend similar things to you.

If Google's devices were powerful enough to do accurate speech recognition locally, do you really think they would choose to do that? Or, at least, choose not to send the text of what you said back to their servers?

Apple is the only one I'd remotely consider trusting, because they make money from selling hardware, not advertising. But we still have Siri turned off on all our Apple devices.
Click to expand...
 
Originally Posted by emg
If Google's devices were powerful enough to do accurate speech recognition locally, do you really think they would choose to do that? Or, at least, choose not to send the text of what you said back to their servers?
I think you misunderstood me.

What I was trying to say was that when speech to text is done locally, then the upload of the resulting text is more difficult to spot due to much smaller data size. For now, they are sending actual audio files which generates more network traffic which is easier to spot.
 
Originally Posted by Quattro Pete
The conversion from speech to text happens in the cloud, not on the device itself, so yes, compressed audio is transmitted to the cloud.

Are we certain about this? I'm not doubting that voice data isn't uploaded, but are we sure that's it's restricted to this by some means?

Not sure if the power argument even makes sense, given that text-to-speech isn't very power hungry, is at a very advanced stage, these devices aren't "dumb terminals" and of course the machine needs to do local deciphering, or how else would it know you're 'summoning its services' without uploading everything to the "cloud" (they're just servers, guys) and having it remotely deciphered?? Doesn't add up.

As with my prior example of facial recognition (and it's advanced state to even have stupid novelty 'filters' widely available for teenage girls to goof around with) accurately placing the cartoon animations over the correct areas of the picture. Do you think that's happening on remote servers or locally on the device? Can you imagine the badnwidth hit if the former was true!? There is no debate that facial recognition is boatloads more device power hungry than simple speech deciphering.
 
Originally Posted by PeterPolyol
Originally Posted by Quattro Pete
The conversion from speech to text happens in the cloud, not on the device itself, so yes, compressed audio is transmitted to the cloud.

Are we certain about this?
That is what I've read based on some quick searches. If anyone has links to some reliable info on this subject, I would love to read more about it.


Quote
[/quote]Not sure if the power argument even makes sense, given that text-to-speech isn't very power hungry,
I don't think text-to-speech is the same as speech-to-text. With text to speech, you are starting with already digital domain which is easy for a machine to work with. With speech to text, you are starting with analog information that I would guess requires more processing power to analyze and convert to meaningful digital data. Maybe I'm wrong.

Quote
is at a very advanced stage, these devices aren't "dumb terminals" and of course the machine needs to do local deciphering, or how else would it know you're 'summoning its services' without uploading everything to the "cloud" (they're just servers, guys) and having it remotely deciphered?? Doesn't add up.
The device is programmed just to spot the "wake" word using local resources. Maybe that's less resource intensive than full-on speech-to-text conversion?
 
Originally Posted by Smokescreen
I don't have any smart TV's, NEST or the like in my domain.

...I am unplugged from the matrix and prefer to remain that way.

Yet you are on the internet, posting here.
 
Originally Posted by Quattro Pete
Originally Posted by Smokescreen
I don't have any smart TV's, NEST or the like in my domain.

...I am unplugged from the matrix and prefer to remain that way.

Yet you are on the internet, posting here.


I love the irony :]
 
Status
Not open for further replies.

Similar threads

OVERKILL
Free VPN apps turn Android phones into proxies
Replies
18
Views
753
Lawnman23
OVERKILL
Chinese hacking group APT31 uses mesh of home routers to disguise attacks
Replies
5
Views
446
OVERKILL
C
Google Fi and spare phones
2 3
Replies
42
Views
1K
Cropduster172
JHZR2
iPhone to USB C transfer Speed?
2 3
Replies
44
Views
1K
y_p_w
Bailes1992
Building A Home Server
Replies
10
Views
419
OVERKILL
Back
Top