Comment by monodeldiablo | Hacker News Reader

Comment by monodeldiablo | Hacker News Reader

Parent Full thread

I've had the misfortune to be present for an in-person demo of a verification service nearly a decade ago. It involved those relationship questions ("Which one of the following people have you not lived with in the last 5 years?") that are incredibly creepy.

I was shocked that they had so much data on me -- I have no debt, no credit cards, no house, no car, no bills, and I had always entered informal rent agreements (I was poor) up to that point -- yet the rep was easily able to list all the places I had resided from college to present, along with a host of off-the-books housemates.

"Where the fsck did you get all this?!" I demanded.

"Have you ever ordered a pizza?"

Turns out, some fast food chains do a brisk business in reselling customer data. I had ordered from Domino's once, but that was enough to link my name to a specific location.

This experience has made me extremely sensitive about the information I give while making a purchase. When I lived in the US, I stopped having food delivered, paid in cash, and never signed up for branded credit cards. I rented informally and, whenever possible, tried to just pay the landlord my share of utilities in cash. Anything to keep a lower profile.

Far more companies than you realize are collecting as much data as possible about you, your habits, and your relations. So I'm afraid your search for a canonical list of data sources is ultimately fruitless. In this new economy, you are always the product.

Do we forget that not long ago everyone's name, address, and phone number was published in the phone book (unless you paid extra for an unlisted number), and hospital admissions were published in the daily newspaper. It used to be unremarkable, and now people shriek "privacy!!" when it's discovered that some mundane detail about their life is not a closely held secret.

No we don't forget.

Those systems were analog and could only be scaled to a certain extent, after which you ran into management overhead.

So in the example given by the OP, dominoes would not be updating your name in the directory.

That would be data that evaporated and never made it to record.

To increase the contrast - whole new types of clustering and analysis are now possible, at the moment a new data point is received.

Indeed, the amount and types of published information hasn't broadly changed, what has changed is the information horizon; the ability to search for information and analyse it. The world hasn't become smaller, our ability to see is much much wider. The information horizon is much further away from zero than it used to be (thanks to the computer's / big data's ability to communicate and analyse), and it's impact on society is much more than just privacy.

This reminds me of the decision on GPS trackers on cars: https://en.wikipedia.org/wiki/United_States_v._Jones_(2012)

Dreeben cited United States v. Knotts as an example where police were allowed to use a device known as a "beeper" that allows the tracking of a car from a short distance away. Chief Justice Roberts distinguished the current case from Knotts, saying that using a beeper still took "a lot of work" whereas a GPS device allows the police to "sit back in the station ... and push a button whenever they want to find out where the car is."

i once worked on a project (about 15 years ago) where we had access to a reverse phone book listing that was generated in part by having a copy of every distinct paper phonebook in the US (also things like school directories) sent overseas to be hand entered (2-3 times for QC). A "nice" feature of it was that the address info was also geocoded so we could answer questions like "Who are my closest 100 neighbors and what are their home phone numbers?". Between that and Google buying the archives of usenet[1] in the same basic timeframe I realized that if it's written down its not likely to stay private for long.

Edit: add citation [1] https://en.wikipedia.org/wiki/Google_Groups#Deja_News

The phonebook was a paper-based system, with significant costs to look up and act on information at scale. You needed the books, the researchers, and the time to find and act on specific numbers.

The numbers weren't widely cross-referenced across multiple other identifier databases: shopping, driving, voting, location to 1m accuracy and 30s time precision.

Every household, small business, mafioso, or political operative didn't have a full copy of the archive, and the ability to deploy it at a moment's notice.

Or in short: scale and costs matter. In fact they dominate all other effects.

Your objection is both meaningless and betrays a profound failure of undrestanding and sympathy.

"Do we forget that not long ago everyone's name, address, and phone number was published in the phone book [...]"

The phone book is an opt-in system, with the ability to remain "unlisted". There's absolutely no such choice, though, from these modern, private data collectors. And they're everywhere.

Consider the following hypothetical: A battered wife flees her husband and takes shelter in a shared house run by a women's organization.

In your phone book world, she didn't need to dramatically alter her existence to stay safe. She could continue to maintain a reasonable subset of her professional and social relationships. She could get a new job in a new city. She could buy groceries. She could pay her own electricity bill. And she definitely didn't need to take her life in her hands to order a cheese pizza.

In this brave new world, though, her ex can track her with the utmost ease. She has to shed nearly every piece of her modern identity -- reconfigure apps, wipe her phone/laptop/tablet, stop using sites that geolocate via IP, and halt nearly all interactions with the modern economy -- at significant social cost.

The sheer number and diversity of data collection points -- and their increasing necessity to participate in the digital economy -- makes opting out orders-of-magnitude more difficult. Yours is a false comparison.

It's likely the locality / ephemerality / repudiability of those identifiers made them less of a big deal in an era when people couldn't instantly look these details up or build a detailed profile of everybody given the right data sources.

Brad de Long addressed this a while back.

https://news.ycombinator.com/item?id=13734768

Wait hospital admissions? Really, why?

I imagine your friends didn't tweet or message a Telegram group about having had an accident back in the day where that was common (whenever that was, I've never seen it). When the telegraph is the fastest method of communication and landline phones are expensive, you might not learn about a friend being in the hospital until word of mouth reaches you.

Not saying this is the one true answer, but I don't find it hard to imagine why.

Temporary postal mail forewards.

USPS NCOA (Change of Address) file is another major data source.

https://www.forbes.com/sites/adamtanner/2013/07/08/how-the-p...

What makes me crazy about having your mail forwarded is that they often don't even forward it to your new address. I know because I once lived with a crazy person and had to move out and they just kept delivering all my mail there despite duly submitting my forwarding address and stressing the importance of it.

I still get mail for the last folks who lived here and they moved out 13 years ago.

But anybody willing to pay the post office for the data can find out your new address.

And then there's how we pay higher rates for first class mail so that junk mail can be cheaper.

It's almost like the government doesn't work for the people.

Thanks for the tip, from the link:

"There is, however, a loophole that keeps data brokers from accessing your updated address. When you fill out the online form to change an address, you can indicate a temporary change that provides six months of forwarding that can then be extended for another six months. That information, unlike the changes marked as permanent, is not included in the master list sold to data brokers."

And the online link for online change of address with the loophole:

https://moversguide.usps.com/icoa/home/icoa-main-flow.do?exe...

Yes, but will they fail to forward the mail as reliably as they fail to forward permanently forwarded mail?

Correct.

(I've mentioned the temporary forwards tip elsewhere in the thread myself.)

Their privacy policy says they don't do this:

https://www.dominos.com/en/#/content/privacy/

That link at the bottom of the email states that they unsubscribe you from the mailing list.

Odd how I always seem to get more spam, but from different companies, after I click one. Almost as though all I've done is confirm that the email address actually exists, and is therefore more valuable to sell on to others.

Perhaps the policy is new. This was about a decade ago.

Regardless, that data is out there now, and no change in privacy policy is going to put that cat back into the bag. Besides, someone else is surely acting as a new source for current data.

It wasn't just my location information, though. They were able to ask detailed questions about my personal social network, before Facebook was ubiquitous and LinkedIn was a thing. They were clearly joining several disparate datasets together to discovery relationships.

You might enjoy the book "how to be invisible" by J. J. Luna

Good to know. Let's all start ordering pizza's, etc under random other names. Create a haystack to hide the needle.

A better strat would be if we all agreed on a single name for everyone to use.

I propose Null or Drop Table.

Why not both? :)

https://xkcd.com/327/

easy to filter out

This would not work, the data is heavily cross-referenced. You would simply have more names (as possible neighbors, or rather co-habitators) in your account.

Reply on news.ycombinator.com