Live data from Hacker News

I still don't think companies serve you ads based on your microphone

simonwillison.net

311–320 of 361 posts

Re: I still don't think companies serve you ads based on your microphone

#311
post #263

Earlier quoted context omitted.

The reason the "phones are listening to you and targeting ads based on what you say" thing is a classic conspiracy theory is that it would take a genuine secret conspiracy to pull it off. Think about how many people would need to be in in the secret and actively lying about it: employees of phone hardware companies, and ad tech engineers, and execs at these companies, and their legal teams, and every contractor they…

I don't think any such conspiracy or secret-keeping is required; one need not attribute any of this to malice which can be attributed to incompetence. There already exists a system, on everyone's phones, which is listening to audio at all times, occasionally activates in response to a wake word and runs search queries and such from it. The point of this system, when used properly, is intended to take a recording of y…

> The occasional false positive response to a wake word that wasn't really a wake word, causing search queries to be run that you didn't intend.

Nobody ever describes that behaviour, though. They don't say "a) we had a conversation, b) our smart speaker suddenly interrupted and gave us some search results, c) I started seeing ads based on the conversation" - b) is never mentioned.

Re: I still don't think companies serve you ads based on your microphone

#312
post #301

The reason I think this isn't happening (other than they have told us it's not happening and technically seems very difficult to do this long without researchers/people figuring it out), is that in my life I've spent millions and millions of dollars on ad budget, we spent so much with twitter they made us custom hoodie, nobody tried to sell this function to me, it's not an enterprise level option, I've seen those, an…

Or, for that matter: if this functionality exists in phones, surely there were some people involved in developing the software and infrastructure to support it. It seems hard to imagine that none of them would have spoken up, even anonymously, to confirm its existence.

Re: I still don't think companies serve you ads based on your microphone

#313
post #291
post #286

The post's arguments can basically be summed up in three categories: 1) "Doing this wouldn't be technically feasible or would require a technical effort wildly out of proportion" 2) "There are lots of psychological biases that lead people to believe something like this happened even if it didn't actually happen" 3) "Apple is such a nice and honest company, they would never do such a thing..." As for 1), there is enou…

Did that tech exist prior to 2017 when this conspiracy theory was already widely circulating?

I honestly couldn't care less if that tech existed back then or not, I care about what's possible today.

Between 2017 and today, there were massive changes, both in technological development but also in mindset: Surveillance capitalism became much more normalized and generally accepted as a standard part of business.

So if you argue that it was a baseless conspiracy theory back then, therefore it automatically must still be a baseless conspiracy theory today, that argument is flawed.

Re: I still don't think companies serve you ads based on your microphone

#314

Earlier quoted context omitted.

I guess my question to this is, what are the estimates on energy required to do this on a continuous timeframe?

Why would they need to listen continuously? Most people don’t speak, and aren’t around people talking to them, for more than a few minutes to hours in a day. You just need to awake every few secs or mins to see if someone is speaking and go back to sleep if they’re not.

Could even switch on when you detect positive social signals implying you are around another person nearby using wifi, Bluetooth, gps, IP address, etc. to ID another device.

They could even pickup or recognize the second voiceprint ID and know it’s your best friend and wake up the audio recognition from ultra power saving mode or whatever. Literally anything is possible to make this is work.

Re: I still don't think companies serve you ads based on your microphone

#315

Earlier quoted context omitted.

I guess my question to this is, what are the estimates on energy required to do this on a continuous timeframe?

Embedded Audio ML engineer here (albeit mostly outside of speech). A modern MEMS microphone uses typically 0.8 mA in full performance mode at 1.8V. Doing basic voice activity detection, which is the first step of a continuous listening pipeline, can be done in under 1 mA. Doing basic keyword spotting is likey doable in 10 mA. But this is only done on the part that the voice activity module triggered on. Lets say that…

This is amazing thanks for doing the math. Didn’t realize the tech was feasibly there already off the shelf. I mean my Apple Watch can detect me saying “Hey Siri” all day with its puny battery.

If big tech isn’t doing this then it sounds like a huge startup idea worth $$$. I hope someone on here in the spirit of HN runs with it and blows the top off this topic once and for all if it’s monetizeable or expose the FAANG patent sharks that come out to play and silence them for infringing on their shady microphone tech.

Re: I still don't think companies serve you ads based on your microphone

#316
post #313
post #291

Earlier quoted context omitted.

Did that tech exist prior to 2017 when this conspiracy theory was already widely circulating?

I honestly couldn't care less if that tech existed back then or not, I care about what's possible today. Between 2017 and today, there were massive changes, both in technological development but also in mindset: Surveillance capitalism became much more normalized and generally accepted as a standard part of business. So if you argue that it was a baseless conspiracy theory back then, therefore it automatically must s…

It's true that it is worth considering that things are possible now that were not possible 7 years ago.

The reason I keep on bringing up 2017 is that it helps demonstrate the psychological angle here. People are absolutely convinced this is happening today. Many of those people were equally convinced back when this was clearly impossible, convinced based on the exact same reasoning (their own personal anecdotal experiences).

So the 2017 thing helps illustrate the lack of credibility of those personal anecdotes, independently of the issue of how possible this might be today compared with the past.

Re: I still don't think companies serve you ads based on your microphone

#317
post #315

Earlier quoted context omitted.

Embedded Audio ML engineer here (albeit mostly outside of speech). A modern MEMS microphone uses typically 0.8 mA in full performance mode at 1.8V. Doing basic voice activity detection, which is the first step of a continuous listening pipeline, can be done in under 1 mA. Doing basic keyword spotting is likey doable in 10 mA. But this is only done on the part that the voice activity module triggered on. Lets say that…

This is amazing thanks for doing the math. Didn’t realize the tech was feasibly there already off the shelf. I mean my Apple Watch can detect me saying “Hey Siri” all day with its puny battery. If big tech isn’t doing this then it sounds like a huge startup idea worth $$$. I hope someone on here in the spirit of HN runs with it and blows the top off this topic once and for all if it’s monetizeable or expose the FAANG…

Hah, that's another great argument against this being a real thing: where are the startup pitches?

If this targeting technique works and is feasible and legal and in demand by advertisers, why isn't there a competitive group of startups all trying to do it better than each other and sell the results?

Now the conspiracy theory has grown to include "dozens of companies compete at this, all of them secretively operating in a marketplace that is entirely invisible to the outside world."

Re: I still don't think companies serve you ads based on your microphone

#318

Earlier quoted context omitted.

"How would it do that? " AI. That is entire premise of 'Nexis' from Yuval Harari. Individualized bot driven surveillance . https://www.theregister.com/2024/09/16/oracle_ai_mass_survei... ""Ellison declares Oracle all-in on AI mass surveillance, says it'll keep everyone in line Cops to citizens will be 'on their best behavior because we're constantly recording and reporting'""

Harari is a shill.

you make a very convincing argument. /s

Re: I still don't think companies serve you ads based on your microphone

#319
post #317
post #315

Earlier quoted context omitted.

This is amazing thanks for doing the math. Didn’t realize the tech was feasibly there already off the shelf. I mean my Apple Watch can detect me saying “Hey Siri” all day with its puny battery. If big tech isn’t doing this then it sounds like a huge startup idea worth $$$. I hope someone on here in the spirit of HN runs with it and blows the top off this topic once and for all if it’s monetizeable or expose the FAANG…

Hah, that's another great argument against this being a real thing: where are the startup pitches? If this targeting technique works and is feasible and legal and in demand by advertisers, why isn't there a competitive group of startups all trying to do it better than each other and sell the results? Now the conspiracy theory has grown to include "dozens of companies compete at this, all of them secretively operating…

Another question that comes to mind now: would this sort of technique run afoul of some wiretapping laws among various states? One is not listing to a wake word to provide a direct response but rather to... idk. just a random thought.

Re: I still don't think companies serve you ads based on your microphone

#320

Earlier quoted context omitted.

I guess my question to this is, what are the estimates on energy required to do this on a continuous timeframe?

Embedded Audio ML engineer here (albeit mostly outside of speech). A modern MEMS microphone uses typically 0.8 mA in full performance mode at 1.8V. Doing basic voice activity detection, which is the first step of a continuous listening pipeline, can be done in under 1 mA. Doing basic keyword spotting is likey doable in 10 mA. But this is only done on the part that the voice activity module triggered on. Lets say that…

Thank you for taking the time to post this informative response. As a sibling comment posted, didn't realize it was so feasible. When posting my original comment, i was thinking orders of magnitude more power would have been needed to facilitate this.
Post reply on HN