I used to work for a mobile advertiser, and we installed hella bloatware on phones.
This idea was floated a couple times but was deemed not very effective cause you’d have to store and process hours and hours of audio data that didn’t tell us much more than just having a week or so of GPS data, your Facebook profile, and your phone IMEI.
It’s pretty easy to see if you’re near a Popeyes and what other IMEIs are connecting to the same tower, extrapolate that to you being near your wife and you and your wife thinking about shit on the Popeyes menu.
Boom targeted ad/video for fried chicken.
The rest is general tech paranoia leading to Apophenia.
There’s no microphones or cameras, it’s just the already gigantic mountain of data anyone who uses a smartphone is constantly broadcasting getting ground through the big data machine that has been the pillar of all tech since the last recession.
And do what? Sentiment analysis on the conversation you were having?
Remember semantically aware models are still fairly new and even they lack the context for a particular field of text. That’s something even the new fancy LLMs struggle with.
Unnecessary when there’s way better targeted models trained on years of data that people willingly send as part of everyday smartphone use.
Sentiment analysis on the conversation you were having?
Among other things, sure. More simply, keyword analysis.
Remember semantically aware models are still fairly new and even they lack the context for a particular field of text.
All of these “models” are useless garbage but it doesn’t stop them from trying to absolutely cram them everywhere they can.
Unnecessary
None of what they do is “necessary”. They could just ask you what your relevant interests are and you could tell them, but they do it anyway. They go to great lengths for any seemingly insignificant amount of data they can get their hands on.
I used to work for a mobile advertiser, and we installed hella bloatware on phones.
This idea was floated a couple times but was deemed not very effective cause you’d have to store and process hours and hours of audio data that didn’t tell us much more than just having a week or so of GPS data, your Facebook profile, and your phone IMEI.
It’s pretty easy to see if you’re near a Popeyes and what other IMEIs are connecting to the same tower, extrapolate that to you being near your wife and you and your wife thinking about shit on the Popeyes menu.
Boom targeted ad/video for fried chicken.
The rest is general tech paranoia leading to Apophenia.
There’s no microphones or cameras, it’s just the already gigantic mountain of data anyone who uses a smartphone is constantly broadcasting getting ground through the big data machine that has been the pillar of all tech since the last recession.
I mean that could be solved as simply as a local transcription service…
And do what? Sentiment analysis on the conversation you were having?
Remember semantically aware models are still fairly new and even they lack the context for a particular field of text. That’s something even the new fancy LLMs struggle with.
Unnecessary when there’s way better targeted models trained on years of data that people willingly send as part of everyday smartphone use.
Among other things, sure. More simply, keyword analysis.
All of these “models” are useless garbage but it doesn’t stop them from trying to absolutely cram them everywhere they can.
None of what they do is “necessary”. They could just ask you what your relevant interests are and you could tell them, but they do it anyway. They go to great lengths for any seemingly insignificant amount of data they can get their hands on.
But wouldnt it be a moot point if I restrict access to GPS for all apps?
How much of that data is from Google/Apple (e.g. Google Maps)?
If you use android google grabs your GPS data regardless, you have to root and disable it.
Apple does the same thing but they didn’t have their pants occupied by third-party network’s fingers like google did until the pixel came out.
Google maps is basically a beacon for AdMob to target you nearly perfectly.
Also using “fine location” in any app grabs the nearby wifi list and sends it to Google/apple if it’s not cached.
Also most ad providers these days have made deals with major networks that let them tell what tower your IMEI pinged off of.
It’s why google tried to push android/ad IDs, way less info for the networks to advertise over, and it also put the tracking in their hands instead.
So graphene os ir a degoolged phone solves the first thing