It's not so much the ads for me - though some recent ones that are clearly scams have been making the rounds here that facebook refuses to do anything about.
It's the whole tracking, data-gathering, and trying to optimize for squeezing the last bit of revenue out of people that I dislike.
That and the stupid amount of bandwidth and compute caused by the ad scripts on every other website. ublock makes the web so much faster, it's hard to believe.
EDIT:
I'm actually subscribed to some e-mail newsletters from certain brands/sectors that I care about, and they regularly deliver personalised ads to a subfolder in my e-mail account. I sometimes even buy things as a result. I don't mind this, because it's opt-in and by consent.
I do mind when facebook tries to infer what kinds of things I might like, which it's generally terrible at and the various "ad preferences" I can set don't seem to make any change.
I find it a bit hard to like the idea of letting professional manipulators try to manipulate me to do things that benefit them. Regardless if they are succesful or not.
I don't like being tracked so that I can be offered personalized ads. I don't like _any_ ads, but at least context ads make sense. Show me ads for waterproof shoes if I'm on a website about what to do with a dog on a rainy day, or for computer parts if I'm looking for instructions on building a computer (I'll still block most of them anyway).
The categories are much to broad to be useful. I've been vegan for about 7 years. The internet thinks I like "food" and shows me ads for meat products all the time. Good to know I'm wasting the ad dollars of companies I think are bad, but I think it's gross and I don't want to see it.
And yet, I still think they can be harmful. Think of someone with alcohol use disorder who recently stopped drinking, or someone with BED who's decided not to keep junk food in the house. You don't think constantly seeing ads for alcohol/junk food would make such a person feel bad or even impede their progress? Why would that be the cost of them opening any website at all?
Or mails sent to the entire households with offers for contraceptives products or abortion assistance.
Or infomercials poping about anti depressants.
True anecdotes. Teenage girl tracked by video surveillance and profiled as being likely interested in contraceptives because she stood near the condom shelves for long minutes without purchase. With a good chance of being pregnant.
(Advertisers could mail to the household, yes. because she provided the supermarket with her address to get groceries delivered once)
A certain messaging app offered by a certain social media platform that mine personal conversations
to profile users down to their emotional states. Those words you type in and send to your confident are put through real time machine learning.
Don't be too surprised you get an ad about chocolate right after you told your date about your favorite ice cream flavor, that's merely creepy. The obsene mental manipulation usually goes unnoticed.
what people object to is the data gathering needed for personalized ads. I don't want all my behaviours and preferences to be collected, inferred and stored. If it can be used for personalized ads, it can be used for other even less desireable purposes.
For me, the answer is tracking. Random ads are annoying, too, but they don't invade my privacy. Personalized ads require the ad network to build a profile of my browsing habits which is something I prefer to have the ability to not grant my consent to.
Because they depend on constant and ever expanding surveillance, and the use of all that data is never limited to just advertising.
Facebook can still show relevant ads without showing personalized ones. For example, if there is a facebook group about car restoration it doesn't take a genius to guess what kinds of ads members might be interested in.
Personalized ads mean they make a ton of assumptions about you using incomplete and inaccurate data. If you actually value advertising as a means of discovery, why would you want your exposure to new things limited to only what marketers think you should be interested in based on stereotypes, or flawed assumptions?
Relevant ads are better because there are fewer assumptions being made. Whatever content you're engaging in dictates what you see, not market research and guesses about who you are.
People often don't like being surveilled relentlessly. Without personalized advertising, the market for all that personal info is significantly reduced, though eg insurers probably still want it.
If you are on your phone showing something to some person, and you get an ad for a new pressure cooking. Well, there it is a nice pressure cooking.
Now if you have been looking for something else that you want to keep private (gay clubs, abortion clinics, or anything embarrassing) then your phone has betrayed you.
There is also a point that if the ad is more useless, the quantity of ads should decrease because advertiser will not find them worth it.
Worse, say a gay club near you was attacked by someone who made an explosive device using a pressure cooker. Your interest in those things (actual or as determined by advertisers) could cause you to be a suspect and/or arrested.
The problem with accepting being under constant surveillance to make advertisers money is that the data is never just used for ads and even if you never show your phone to another living soul that data never goes away and can end up in the hands of just about anybody.
Does not work as imagined by most, and is easily detected by sites. So basically just a convenient way to have your cookies and other site storage deleted. But (meanwhile) basically useless because of other ways of 'fingerprinting' client browsers.
Hating ads in general, it's logical to hate them succeed even more. It's the positive feedback loop that leads to force-fed dystopia IMO. It's a temporary remedy, but I do "bad actor" behavior, once I get shown a good and relevant ad, I am banning the source and interact with the irrelevant ones instead.
I think only a vocal minority hate them so much (it is a majority here).
But that's because it is creepy, if the targeting is too accurate, it feels like you are being watched. Which is true, but a little bit ironic on Facebook and Instagram where people have no problem exposing their entire life to everyone.
Too much private information collected for that ads will be catastrophic sooner or later.
I don’t think that people got the real wake-up call yet. For example imagine someone getting addresses of all the Jews in town and doing something nasty. Before that you needed a lot of people to gather info for such an operation, now the information part will be trivial, but there is still an IRL part. Or getting the names of all people with severe peanut allergy and adverting them something with undisclosed peanut consents. I think there just have not been a sufficiently motivated actor to cause hard beyond making you buy more, selling scams, paying more through individualized pricing or discriminate based on some parameters.
Because they go with mass corporate surveillance. As soon as a company starts personalizing ads with some data, they’re going to be under financial pressure to personalize ads with all the data they can get.
I take it that it's the amount of surveillance required to personalize the ads. I certainly don't like the idea of shitloads of data being collected about me.
Now I do like personalized ads and I get insane ones even though I'm anonymizing my tracks more than most.
For example I do get personalized ads trying to sell me... Private jets!
I mean: I'm maybe upper middle class but there's no way I've got the money to buy a x million private jet.
Yet I get the ads for them Falcons and Gulfstreams.
The default position should be that advertising is bad for you. Therefore more effective advertising is worse for you. Personalized ads are more effective, thus they are worse for you.
Random spam is just random spam. Personalized ads are based on behavioural data that they have harvested from people, it seems that makes people feel iffy about it. Seems some do care about privacy after all.
The entire point is that you can't opt out. The sites are legally required to let you, but they don't. For example, by claiming that the tracking is OK under "legitimate interest" and doesn't require your consent, which is what Facebook has done so far.
Or by giving you the "choice" of paying an absurd amount of money or "freely" consenting to them harvesting all your data, which is what Facebook has already announced is the next step they'll take. Whether this is legal or not under GDPR is hotly debated. While that debate is running, Facebook will happily continue (if you pay to "opt out" they'll make even more money off you, so win-win for them). If the debate concludes that it was illegal, they'll either get away with it because "it was unclear so we can't punish them", or they won't, and they'll pay a fine that at that point will represent a small tax on the additional profits they made through this practice.
Either way, they win, consumers lose, because GDPR enforcement takes 5 years per round and multiple rounds...
Not at the moment. We're considering making it source available though. What are your thoughts on this one? It's still early so we're happy to adjust some things given demand!
We're a small team of engineers that spent waay too much time setting up user-triggered distributed workflows with K8s when working on our previous product, even though conceptually we knew exactly what our system is supposed to do [1].
And so we decided to build Multinode. Its goal is to eliminate the hassle of setting up compute infrastructure so you can build arbitrarily complex compute without leaving the comfort of your local Python environment.
You can use it to set up distributed async workflows, continuously running daemons, APIs, etc., with a help of only 5 additional annotations.
TECHNICAL DETAILS
Internally, all the infra is on AWS. All the compute runs on ECS with Fargate using controllers similar to those used by Kubernetes. Unfortunately, Fargate has pretty annoying resource limitations (hence https://multinode.dev/docs/resources/cpus-gpus-memory) so we will likely port it to ECS on EC2s or just straight to K8s at some point. The task registry and multinode's dict is running on Amazon Aurora. We use Python for the actual product and docs run on Firebase with next.js using one of the official Tailwind templates (I'd really recommend them!).
FEEDBACK NEEDED
In its current version, the entirety of your cloud compute would be in Multinode. But the actual #1 reason we decided to build it was these user-triggered autoscaled workflows [1]. We're now playing with an idea of dropping most of the features to focus on solving this core problem in such a way that it's easy to integrate with your existing infrastructure [2] instead of having to build everything in Multinode. Something like Airflow or Dagster but with emphasis on user-triggered workflows with very fast autoscaling and topologies defined at runtime. Do you think we should go with this change or keep Multinode as it is now?
Also, apologies for closed alpha at the moment, we're still figuring out how to put automatic pricing or quotas in place and so opening it up to the public would bankrupt us. In practice it will be some markup on our compute cost. Any advice on what is usually reasonable here?
Any thoughts and feedback on those 2 points would be very appreciated!
[1]: For curious, it was something similar to our chess example: https://multinode.dev/docs/core-concepts/job#distributing-a-...
[2]: Ideally, both compute and storage would be provisioned in your own AWS VPC with only the control plane running on our infra for maximum privacy and security.
EDIT ---
Ok, I get it now. Personalized ads = surveillance. Fair enough.
Doesn't the whole GDPR already cover it though? You can opt out of the surveillance.