Privacy Labels and Lookalike Audiences | Hacker News Reader

Privacy Labels and Lookalike Audiences

stratechery.com

42 pointsaskl5617 comments

I completely disagree with this article. It is almost like he assumes that data harvesting at the level of Facebook and Google is normal and should be accepted. Quoting from the article, "Computers emit data as a matter of course, and the Internet makes the transfer of that data free. To strive for a world without the generation or capture of data is to fight against the very nature of technology."

No, my device generates data and I should be allowed to know who is gathering that data. If the application on my phone is doing things like trying to access my pictures or Bluetooth settings, I should know and be allowed to block it.

I have a right to privacy. I appreciate that Apple is making it slightly harder for companies like Facebook and Google to harvest my data.

This article isn’t about access to photos or Bluetooth.

I don't know - on the other side of the coin imagine if no data was collected. It would be near impossible for companies to do any sort of troubleshooting or QA.

Taking it a step further, it'd also be significantly more difficult for them to understand their audience and users... which could actually make the experience worse for those users.

Even in the more nefarious case of ad-tech, I'd personally prefer more relevant ads than generic non-relevant ones.

I'm not saying all data collection is fine and well intentioned, but I also don't think it is necessarily as zero-sum as people think.

I do think when data gets licensed or shared w/ 3rd parties it should be more clear how it gets used.

> Even in the more nefarious case of ad-tech, I'd personally prefer more relevant ads than generic non-relevant ones.

We don't need surveillance infrastructure to achieve this. Advertisers can simply place their ads on pages that are relevant to their audience.

> I don't know - on the other side of the coin imagine if no data was collected. It would be near impossible for companies to do any sort of troubleshooting or QA.

Good. Companies got by in the pre-internet era by running supervised QA tests with customers in person, or using special devices modified to allow surveillance. There's no reason why this cannot be brought back.

Nokia phones used to have a feature called "AutoSMS" which would silently send back crash reports and various usage metrics. Of course this feature was never included in production releases, only for field/user testing. After product launch, feedback was gathered by surveys and crash reports from service centers. Obviously phones have always had the capability to send covert data, but back then it would have been utterly unacceptable. How times & business models have changed :(

>Even in the more nefarious case of ad-tech, I'd personally prefer more relevant ads than generic non-relevant ones.

Personally, i'm the opposite. Every time I see an obviously targetted ad based on something i've done, watched, said in a text conversation, my emails, etc. It creeps me the hell out.

> “ Facebook data is valuable only to the extent that it is usable on Facebook”

This is an incredibly wrong and misleading section. That data can be processed through data fusion centers and then joined against other data aggregators - all your physical location checkins from Yelp or Foursquare, all your grocery purchase history, what activities you booked on your last vacation.

I used to work for a well-known stock photography company and the company collected huge amount of data - browser signatures, upload geolocations, login locations, in addition to the actual images and customer behavior within the product.

Most of it was completely unused and one of the senior VPs of engineering was fond of saying we were making “a bet on data.” They had no immediate use for the data, but felt strongly that the larger and more comprehensive they could make an in-house data warehouse, augmented with as many types of data as we could collect, the more valuable it would eventually be in the future for use cases or third party sales that just hadn’t been thought of yet.

They are still doing this today with no uses of that data yet (or private uses selling to third parties like Facebook).

I can’t believe Stratechery would have an article so completely wrong on the absolute basics of data hoarding.

The article also claims the captured data will never be leaked:

> [Apple's privacy ad is] truly scandalous, but not for the reason Apple wants you to think it is: the way in which this ad depicts how your information is used — literally broadcasting your web browsing and searches and private conversations to those around you — is so misleading that it is hard to call it anything but misinformation.

It's like they have never seen any of the huge data leaks that occurred over the past few years, or that whole NSA data collection system. There is no such thing as computer security in 2020. If there is data on a networked computer, it should be treated as effectively public.

The article describes how privacy could be compatible with data collection in some perfect fantasy world, not the real one we inhabit.

> To strive for a world without the generation or capture of data is to fight against the very nature of technology.

How is "the very nature of technology" to tell every site on the planet my GPU?

I get the worries about the anticompetitive nature of regulation. (Goes back to George Stigler:"regulation is purchased by the industry".) I'd like to see more evidence that small businesses need targeted advertising to survive.

Also, I'm not sure that the opposition between "small businesses" and "giants" works. The giants are mostly platforms for the small businesses to work off. Amazon is a marketplace too. There's no bright line between "integrated giant" and "platform for micro-businesses".

Someone should tell the author Facebook do a lot more than simple lookalike audiences with the data they collect.

> What makes that Apple advertisement so misleading is the level of individuality it implies in terms of data collection and application. Individuality is a real problem when it comes to data collection

I guess the author has never come across this before https://www.facebook.com/off_facebook_activity

The "Off-Facebook Activity" page is reasonably new however the data it shows goes back years, suggesting the data collection itself has been going on for a long time and Facebook has now become generous enough to let you know about it.

There's reason to believe even more data is being collected & processed behind the scenes but they don't tell you (yet?).

This stuff has been collected for as long as Facebook have offered pixels, widgets and SDK's for webmasters and devs to use on their sites/apps.

This was a very unconvincing argument.

1. Apple's analogy with nutrition labels was strictly how transparency is a good thing. The food pyramid add-on analogy felt like a straw man argument.

2. The Apple video of someone broadcasting their information to others isn't misinformation. It seems inevitable that these systems will eventually be compromised or exploited. Facebook has gotten in trouble with this a couple times through exploits and people misusing app permissions to capture data.

3. The article honestly read like shilling. They didn't engage the strongest arguments against their position. They made it sound like people's issue with unlimited data collection is strictly due to their misunderstanding of how Facebook does targeted advertising. They even employed a "think of the children" style argument.

I wish apple applied these privacy nutrition labels to their own products. There would be a lot of eye raising activity.

I normally am a big fan of stratechery but I couldn't disagree more, especially in the reality of how data is collected and SOLD.

Analytics packages do tie things to individual consumers and while that may not be useful to the company compiling the data, it does get sold that way. And yes, I can tell you there are companies that crawl through individual level data and use it for everything from very targeted marketing to basically scams and illegal uses.

Also the issue comes that many companies have garbage security. Even if they don't intend to abuse or sell the data, it will almost certainly walk off and become public. All the good intent of stopping child exploitation won't fix your world open S3 bucket. Oh and once it's leaked? Yeah, it'll get packaged and resold.

The idea that we need to invade privacy to stop crime is hilarious. We stopped crime before the internet and we will continue to stop it. One thing the article gets wrong is that it's automated scanning that stops child pornography - while that does limit reuploads, most pedophile rings producing and selling content are busted by.. detective work and flipping insiders. Yeah it's great to stop reuploads but that doesn't stop the original crime - arresting the people creating and selling content does.

Data once collected can't be uncollected. The genie doesn't go back into the bottle. Already we're seeing the EU try to crack down on freedom of speech - and right now that might be against undesirables like Nazis and anti-semites, but it will be turned against others like it always has before. Already we've seen people protesting against police be targeted through this kind of data targeting, and there was that whole Belarus/Telegram incident.

Reply on news.ycombinator.com