Hacker News Reader

Comment by matheusmoreira | Hacker News Reader

> It's a bit like the police asking the library who checked out a specific book.

And you have no problem with this? Government and police looking into and judging you based on what you read is one step away from policing wrongthink. You can't learn anything deemed dangerous anymore without being arrested. Can't look up how a bomb or sarin gas works without SWAT coming down on you because you're obviously a terrorist.

This government collusion with Google is the same thing. Government and police show up and say "find me a list of people who searched for dangerous knowledge". This is normal now, without even the trepidation depicted in films like Se7en.

View on HN

They said “specific” book, which I take to mean “we found this book at a crime scene, can you tell us who checked it out?” Not, “I want to know everyone who checked out war and peace this year.”

But after thinking about it, I think they meant the former and you meant the latter. But to the library, it’s the exact same thing. A library doesn’t (traditionally) track each and every book, so if it’s a popular book, it could have been checked out by any number of people.

I think this is fine, as long as courts work to keep the scope small and laws allow people to do whatever they want without hindering freedom.

Libraries track books individually. If you and I each check out a copy of War and Peace, and you turn yours in on time but I don’t, they know who to pester.

I think there’s a continuum from the cops being able to say “we found this here book at a crime scene, who had it checked out?”, which should clearly be allowed, to “the security footage shows a copy of War and Peace with the libraries sticker on it in the crooks hand, who had it checked out”, which should probably be allowed, to “we need to know everybody who checked out The Anarchists Cookbook in the last six months because the crook used an IED thats described in there”, which probably should be allowed.

I hope you forgot a “not” in your last example!

Indeed!

> If you and I each check out a copy of War and Peace, and you turn yours in on time but I don’t, they know who to pester.

They track checkouts. But back in the day, we could switch checkout cards in the cover of the book and return them and the library would never know the difference. With modern computers and barcodes, I do believe the individual books are tracked.

I think the library warranty thing depends. If the government wants to find out who checked out the Communist Manifesto so they can jail them for thought crime - that's bad. If the police want to know who checked out this specific copy of Harry Potter because they found it at the scene of a murder, that's good.

Sure, I might've just dropped my copy of Harry Potter near a murder as I innocently passed by. And yeah, I might've been looking up the victim's name and address in Google purely out of curiosity right before they got murdered. Neither the Google search nor the book are dispositive but they fit into a broader context which could be used to convict or exonerate me.

Ultimately we need laws or judges to set reasonable limits on powers that could be used for good to make sure they aren't also overused for ill.

There's an in between situation where someone is looking for correlations with a crime, but they are simply searching so much data, that they inevitably get a huge number of very close, but invalid matches.

This tends to be a problem more and more with ever increasing databases that are convenient to access.

People have intuitions for coincidences that are not appropriate for vast amounts of data.

This is a really great insight, and demonstrates where the "library book" example is lacking.

Since Google is so vast, I can find a page of search hits to support any particular belief I might have, no matter how wild.

Things like this make me think of the movie Gattaca. Mild Spoilers After the murder, they just go around vacuuming up any and all organic material to test for DNA. They get a hit from someone who “isn’t supposed to be there” and immediately focus the entire investigation on that person. It fits the very real pattern of using science to just confirm the human biases held by the people using it. If police can sweep up and search through a mountain of data, it is going to be the Muslim immigrant hit they focus on over the white middle manager. It will be the person struggling off and on with homelessness that gets arrested because they matched a pattern and didn’t have an alibi, over the suburban mom.

That if you give police a list of random correlations, they will focus on “those people” that are assumed to be more criminal and usually have less means and ability to fight back.

> And you have no problem with this?

It was only a problem when George W Bush wanted to do it. :-p

Reply on news.ycombinator.com