Hacker News Reader

Comment by Johnny555 | Hacker News Reader

I'd blame this on confirmation bias - there must be dozens if not hundreds of times you use or mention a product every day that you don't notice ads for, but the rare time you do, it reinforces the belief that they are reacting to what you said.

View on HN

It could very well be, but do you know what kind of evidence you would have to see to conclude that FB is listening? Otherwise you're just making a fully general argument against the possibility of this (or any such scheme) happening.

Side anecdote: One weekday after I vacationed in Tahoe, I saw a BART (subway) ad for Tahoe, and I was like, "oh, great , probably because I just came back from ... wait, that's not possible!"

but do you know what kind of evidence you would have to see to conclude that FB is listening?

Any one of:

* OS-level confirmation (permission entitlements, cpu usage, etc)

* Packets resembling sound data being caught in flight.

* An internal leak of the method they'd be using to do so from one of three of the largest tech employers. Not even Apple can keep their secrets secret, and they're probably the most paranoid tech company in existence.

You know, literally anything concrete, rather than evidence-free accusations based on fallible memory.

So far, not one bit of these instances can't be explained by a combination of Baader-Meinhof and confirmation bias, with a mix of plenty of non-audio data that Facebook no doubt has. People are so willing to paint FB as this boogeyman that they're disregarding basic logic.

So do I understand you correctly that you’re categorically excluding any kind of experimental evidence, no matter how well controlled or rigorous? That, so long as there is no breach of the source of the technique, you can’t be convinced?

I understand that any one person’s anecdotes are weak evidence, but your comments are going much further and claiming that such tests can never be evidence, even though much scientific knowledge is similarly obtained.

Experimental evidence is fine, given an actual experimental protocol. So far, all we have are weak anecdotes that don't come close to proving the assertion.

>* Packets resembling sound data being caught in flight.

Not even necessary. You could do keyword recognition on the device itself, pushing a list of keyword<->waveform maps, and sending an indicator when they're recognized.

I don't have a FB account, but my wife's account absolutely delivers her ads intended for me. I've not been able to isolate exactly how it is doing it, but the dragnet is absolutely wide enough to capture a household.

She's received ads for Civic Type R (she hates cars), Senior Java Developer (she works in a totally unrelated field), cooking tools (I do all the cooking), and tons of other things. It creeps me out.

I’m guessing you and your wife share a home and an IP address, and Facebook is using that to associate “your” ads with your wife’s FB account.

Edit: How Facebook would get your browsing data, even if you’ve disabled things like ads and those FB like buttons (like I assume most HN posters would), is beyond my wild speculation.

If they have positional data they could do a correlation of location speed, along time time of week. If they constantly see you going to the same places on the weekends, that might be a clue, especially if in close proximity. Especially if traveling at the exact same high speed in close proximity (Driving there). GPS, at the high range, can tell location within meters.

And none of this would have to happen while you were talking about cars, or anything else. Just enough times to make the correlation, and with data that could have been collected months to years ago. Google Now did that for my commute from work to home.

I think it is more likely Facebook is using location data to create edges on a shadow social network. We just bleed metadata.

My current suspicion is they think my phone is hers due to IP. She doesn't use FB mobile at all and I use ad-block + NoScript on my computer, but not on my phone.

I've picked two different, random topics (boats & umbrellas) to occasionally search for on my phone and computer respectively. So that should help me figure out what the source is.

Why don’t you run an experiment, say something totally random that you know you’ve never searched and is out of the range of normal interests you have and see what happens. I have done that on more than one occasion and it confirms we are being recorded. It’s not just Facebook.

I did that experiment this morning -- before work at home and on the way to work, my wife and I were discussing getting a new car, and spoke about a particular car brand that a coworker just purchased.

I just asked her to check her Facebook and she doesn't see any car ads. I checked too but didn't see any ads for cars or that brand, but I don't have the Facebook app on my phone (she does) and our Facebook profiles aren't strongly linked (i.e. she's not listed as my "wife", we just friend each other). She uses her phone for navigation while driving, so it was in a position to clearly hear us.

I haven't done any research on that car brand, but I suspect that once I do a Google search, then the ads will start flooding in.

So maybe this is confirmation bias in the other direction, but I don't see any evidence that Amazon Alexa, Facebook, or Google Assistant are spying on us. Though it could just mean that this particular carmaker doesn't purchase ads based on keyword spying

For experiments like this, use items that have high CPM. Gold, silver, niche personals and preserved food are great canaries.

Try doing different things. I don’t think name brand vendors do pervasive audio surveillance. I do think they broaden the scope of your intents. Use GBoard dictation or similar tools to write. Write stuff down in different contexts. Use apps in different ways.

Amazon and Facebook share in near real time. Anything you do in a consumer Amazon property is feeding context to FB.

Sorry, I didn't mean to imply that I wanted to do this test, I just happened to do it by accident this morning -- using a very similar term as the writer of the parent post, he said he mentioned Scion, I mentioned another major car brand that starts with S.

But you aren't being recorded or listened to. Countless security researchers have dug into this problem and can tell from the traffic that there is audio being recorded. Furthermore, if facebook is accessing the microphone without permissions, it would be through an unknown security vulnerability and would be against the terms of service with both android and ios.

You are just not noticing these ads until it is something you have called out. You would never have given a second thought to this totally random item otherwise.

I have tried the same test as you multiple times just for kicks and have never found it to be confirmed.

Well you surely can't deny that SOMEONE is listening? Google or Facebook, do some tests and you will find it's true. Confirmation bias is a possibility but from what I have seen and experienced I just can't believe that is always the case.

Reply on news.ycombinator.com