Little Things That Irk You: The Hateful 08

[QualityChimp]

The worst thing about super-accurate algorithms giving you suggestions is that you no longer get those random Wish adverts trying to sell you cat saddles, artificial ears or a trombone with a gynaecological mouthpiece.

[Dogg Thang]

The worst thing about super-accurate algorithms giving you suggestions is that you no longer get those random Wish adverts trying to sell you cat saddles, artificial ears or a trombone with a gynaecological mouthpiece.

Well I will always have the pleasure of Amazon trying to sell me random junk or more of what I just bought and so on. For a company with so much info on our buying habits, they seem to have no idea what to do with it.

[wakka]

I feel like we can move past this idea precisely because Facebook have been caught on transgressions on more than one occassion and that didn't stop other transgressions. In your words, FB have shown themselves to be sufficiently amoral and also not all that afraid of the consequences.

I guess it's not just that I think it would seem especially damaging if they did it. It's more that it would be damaging, plus someone would definitely figure it out.

It feels like something that someone would have identified that it's definitively going on before now. I'm no software engineer but wouldn't an app doing always-on microphone recording be identifiable or recognisable in some way?

Whereas sneakier methods of getting the same information - like recording the WiFi network you're connected to, which seems innocuous enough, but then matching that data server-side with other users connected to the same network to identify that you're with them - seem like they would be much harder for anyone to find out about.

I don't know anything more than you, obviously - this is all just my instinctive feeling, so it's essentially worthless.

[Dogg Thang]

Well they definitely track your wifi network and those close to you - that's in their documentation and it's for your safety apparently. As for the mic, as Brad says, it has to be difficult on iPhone without permission. No idea about Android. But then presumably they are linking lots of pieces of info so can they also add stuff from your desktop etc? They are everywhere.

[Brad]

Yep, it is. But where it gets more terrifying for me is not the relevant ads, but the irrelevant ones - because, as with a diabetes tracker, nothing in my info would point to that ad but there is very obvious trigger for it. And it's those ads that are usually the most noticeable and usually the trigger seems to be around 12-18 hours before the ad is served up which often makes them very easy to identify but, in some cases, very hard to figure out how FB got it. And again, I can't say for sure how they are doing it but, having seen it over the course of many years now, I can safely say it goes beyond cookie tracking, as in just tracking my own cookies.

But yes, it is terrifying and it's funny because it's ones like this that stand out, ones where it feels like there is no way FB should have had that info and yet what I pretty much take as normal now is the everyday obvious tracking that FB do that they shouldn't do. Search for something on Twitter, just as one example, 24 hours later you might find a related ad popping up on FB. At what point is FB intercepting that information? I don't know but it doesn't seem like FB should have access to even that kind of tracking. However they are doing it, even their day to day normal tracking seems horribly invasive.

I don't pretend to know their algorithms, and you're right I should think; the diabetes thing was probably not a coincidence. The colleague you were chatting to, are you connected on FB or Instagram as well? I'll assume you are and what follows is totally wrong if you aren't but if they were talking to you about diabetes you might assume that the topic came up in their digital footprint as well at some point, so they get targeted for diabetes stuff and, by association, so do you. The odd thing is the more you cover your tracks the more likely weird stuff like this is to happen. If they can't target your interests directly because they don't know them then they had to advertise SOMETHING to you so they go further afield. I often use a VPN and usually go via a German server. Now some of the podcasts I listen to have German adverts, even though I wasn't on the VPN when I downloaded them or when I listened to them.

All these companies use every bit of data available to them. The WiFI MAC address spoofing does hide you to an extent but that depends on how often your device changes your MAC address. In addition every device has a device type/manufacturer identifier that is easily obtainable from unencrypted management packets. So, if there's only one device on a network with your device type identifier then some AI software could make the assumption that it's the same device it saw earlier, even though the MAC address changed. They can use all the combined data to generate a likelihood that they're seeing the same device and therefore the same person. They don't need 100% accuracy, they just need to be correct now and then.

[wakka]

But then presumably they are linking lots of pieces of info so can they also add stuff from your desktop etc?

I think that's pretty much the nub of the matter. The cleverness/invasiveness/outright evilness is in how they are linking disparate pieces of information to try to calculate what advertising you'll respond to.

The diabetes thing feels like they are connecting you to the friend you were talking to, who had had presumably done lots of research around these products. So Facebook thought it was worth testing whether you might be interested in them, too.

How they are connecting you to that person, I don't know. But I reckon that's what it is, rather than mic tapping.

EDIT:
[MENTION=9465]Brad[/MENTION], the connection between Dogg and his friend must go beyond being Facebook friends/Instagram followers of each other. That information is of pretty limited use as people are friends or followers of hundreds/thousands of people often.

They must be able to draw a much more direct connection - it feels like they were potentially able to identify that Dogg+friend were in the same location for an extended period.

[Brad]

@wakka maybe. Not necessarily though. Although if they were both on the same Wifi network, or had both allowed Facebook or Whatsapp or Instagram to access their location then they could definitely correlate that. @Dogg Thang do you ever see any ads that are totally irrelevant to you or do 100% of them seem to be targeted?

[Dogg Thang]

Yep, the connection would have to have been more direct. As for this particular instance, the person I was talking to doesn't have diabetes and wouldn't be in that targeted group either but it is entirely possible they did a search for related stuff after we chatted - a movie trailer reference is the most likely thing. A trailer for a new movie showed a kid with a diabetes monitor on their arm and that's how it came up and she may have searched to see that.

As for their accuracy not needing to be 100%, that's where there is a massive difference between FB products and the others (like Google). Because for me, there has been a noticeable refining of that info to the point where I can see why pretty much every ad on Instagram is served up to me. Their info isn't 100% accurate but it's insanely close, which is why the anomalies stand out. And they are never random. Like I say, there is always a trigger. And yes, you might not spot that unless the info hits a point where it's actually getting close to 100%.

Edit: Brad, no. Not on Insta. I can identify why every ad on Insta is served up to me. It didn't used to be the case but they got better and better over the years.

[Asura]

Something funny about this if you ever use a VPN on a fresh browser. Due to how those work, the algorithms have literally no idea who you are, so you get ads for all sorts of random things, usually local to where-ever the host is; so if it's in Indiana, USA, you'll get loads of stuff local to that lcoation because it's the only consistent thing they know.

[QualityChimp]

I think we're all in agreement that the browsers know something about us, and we're debating how much, but this conversation stemmed from my original point that the tech giants have a myriad of ways to target us for adverts, but seem to have very little control over racism on their platforms:

https://www.theguardian.com/world/2021/jul/14/social-networks-anti-racism-policies-belied-by-users-experience

[Dogg Thang]

I think the main element there is nothing to do with a lack of tech or lack of ability but instead a lack of desire.

[Asura]

I think the main element there is nothing to do with a lack of tech or lack of ability but instead a lack of desire.

It's also partially related to how, I think, some of those creating the platforms were too naive/positive in their outlook at their inception.

I can't help but think that part of this is because human beings, unfortunately, are *****. Too many of us are easily manipulated by anger and negative emotions. Like you can post a headline about how someone has donated their bone marrow to save a kid, and another about how a teenager has asked for gender-neutral toilets at their school, and we all know which of those two things is going to blow up.

The algorithms which push these things are, to some degree, neutral. They're heuristic processes which spot engagement and adjust visibility to maximise it. I don't think they're intended to be homophobic or racist; it's just that a morally neutral self-reinforcing algorithm is going to amplify negative stuff, because of the human element that shunts them.

That means if they're going to try and enforce standards about racism, sexism etc. on their platforms, platform holders need to make a conscious decision that their algorithms will not amplify things in a neutral way, but instead have to filter and only amplify socially-conscious things. That's a big step for them, and back in ~2012, to a small degree, I could empathise with their reluctance to do so...

... but that empathy has entirely eroded in 2021 when Zuckerberg and his ilk have repeatedly talked about how their service is changing the world. Like they know that like it or not, they're the new Murdoch; they wield an enormous amount of power over the thoughts and minds of humanity, and it's time of them to **** or get off the pot; in practical terms it might be time to accept that such a thing can't be owned by a private firm that is run for its own self-interest.

[wheelaa]

https://www.bbc.co.uk/news/uk-57896088

""I'm writing this not as the prince I was born but as the man I have become.""

I managed to snag a preview copy...it reads, in its entirety, as follows*.

"I was a born a thick ****. I'm still a thick ****. I will always be a thick ****."

*Quite possibly not 100% true...

[fuse]

We have recently gained some noisy neighbours; not the usual loud-music type, rather "we have all doors and windows open and shout and scream and laugh ten times louder than normal humans do, throughout the night". We originally thought it was to do with the football and let it lie, but it's carried on since. Because its flats and we don't know the address, the council's noise team has said they'll have to come to us and hear it in order to do anything, but given they close up at 1am (it usually starts later than this), I feel like we're going to get nowhere. I'm a pretty deep sleeper, but my girlfriend is getting woken up by them almost every night. Not scouting for solutions, just sharing an irk.

[fishbowlhead]

Call the police. Failing that water balloons filled with butyric acid through their windows, its a chemical that comes from pineapples and the smell will NEVER come out the flat, EVER, they’ll just move out.

Your welcome.