Hacker News Reader

Comment by tlrobinson | Hacker News Reader

“That is kind of weird,” she laughed. “I’m still not so sure this isn’t just coincidence. I don’t think Facebook is really listening to our conversations.”

This is a terrible article. There's an easy way to figure out if it's a coincidence... repeat the experiment.

View on HN

99% sure this is not a coincidence, here's my theory: people's IRL conversations can be pretty reliably predicted by their Facebook conversations, so targeted ads based on messages tend to be surprisingly relevant to the next thing they talk about IRL.

Example: my girlfriend was given a box of Flonase by her parents and immediately got a Facebook ad for it. She had not mentioned Flonase anywhere on Facebook ever. Spoopy eavesdropping? No --- she did talk on Facebook about her ENT appointment that morning. And what products might be relevant for someone going to an ENT doctor? Allergy medication.

And in cases like that, it's very likely that your two recommendations from different sources will happen to match --- what brand of medication are your parents likely to recommend? A popular one that spends a lot of $ on advertising. What brand is likely to buy ads? A popular one that spends a lot of $ on advertising.

See this seems reasonable.. however I was once playing Battlefield Hardline raving about how awesome the camera's were for whatever I was doing and was then presented with security camera ads shortly after while browsing Facebook on my desktop. I have no interest or need for security cameras or even any reason to source information about them online. I double checked my web history and couldn't find anything related either, plus I had not had any conversations about cameras on Facebook. I remarked about how creepy it was over Team Speak and had 2 other people comment about similar experiences recently (at that time) as well. They claimed the only device they had near them was a phone at the time and the conversation was spontaneous.

Sample size of 3. Evidence collected through anecdotes. I'm convinced.

See frequency illusion.

https://en.wikipedia.org/wiki/List_of_cognitive_biases#Frequ...

This seems like a plausible alternative explanation.

An experiment could easily be ran where a new Facebook account is created without any posts or messages, and the Facebook App is installed on a mobile device. Spend a few days using language that is easy to advertise to in proximity of the device, and see what happens.

> 99% sure this is not a coincidence, here's my theory: people's IRL conversations can be pretty reliably predicted by their Facebook conversations, so targeted ads based on messages tend to be surprisingly relevant to the next thing they talk about IRL.

That may be true, but my friends and I once did this with a made-up phrase about some housewares (think 'pots and pans' or something similar): we repeated the phrase over and over in discussion that evening — and sure enough, an ad popped up on Facebook.

Not gonna install it, just not going to. And I'm beginning to think I should politely ask people who do have it installed to go elsewhere.

I agree that the nature of the article is bad but still I would not be surprised to hear that Facebook is listening to phone conversations. That's why I don't use it. :P

I work for Facebook. This does not happen. If it does, I can assure you that one of the many thousands of engineers will speak out.

What makes you so sure that many thousands of Facebook engineers would be aware of it?

If something like this is going on, only a few select employees know about it. Facebook is not dumb.

When it comes to "speaking out"... NSA started surveillance projects long before Edward Snowden was employed. There will always be people who will develop things like this and who will not speak out. It's just a matter of correctly selecting these few people.

Repeating the experiment wouldn't disprove the assertion. What you would be testing is "(a) Facebook is really listening to our conversations and (b) consistently responds to it in this way". A negative second test could never disprove (a), however a positive test would confirm both (a) and (b).

Reply on news.ycombinator.com