Hacker News Reader

Comment by shady-lady | Hacker News Reader

I personally don't really care about what information they collect. I do hugely care who they give that information to & what they do with it.

Haven't heard anything about Google wholesale handing all my data over to anybody who clicks a couple buttons to sign up to their developer program though.

I'm guessing Facebook has the technical expertise to allow 3rd parties to run aggregate (non identifying) queries on the consenting users' data on servers they control and only allowing certain aggregated data, as well as limiting the number and type of queries 3rd parties are running.

Guess it's easier to just let anybody have at it to prove how valuable their platform/data is.

View on HN

Google keeps their data private because they've already secured their advertising/search territory. Disclosing profiles would only undermine their competitive advantage.

It seems like Facebook gave away that information in hopes of developing a more engaging walled garden with 3rd party help, and perhaps some naivety.

> I'm guessing Facebook has the technical expertise to allow 3rd parties to run aggregate (non identifying) queries on the consenting users' data on servers they control and only allowing certain aggregated data, as well as limiting the number and type of queries 3rd parties are running.

This is a very challenging problem - multiple "non-identifying" queries can actually identify individuals. Differential privacy is the best solution so far, however it's still challenging to guarantee privacy without lowering quality far below competitors that don't care about consumers.

> Google keeps their data private because they've already secured their advertising/search territory.

That is one possible self-interested reason for them to keep the data private. They could also just care deeply about guarding user data they collect. This could be for either selfish or unselfish reasons -- it's both good business and moral. It's very hard to tell from the outside, because their actions would look largely the same either way.

Personally I think it's pretty naive to believe any company does any thing for a single reason. There are a constellation of reasons, some more important than others.

> multiple "non-identifying" queries can actually identify individuals.

Zero challenges from me on this. I 100% agree. However, it is doable. It just takes effort.

> I personally don't really care about what information they collect

If the government said "let's make a database of every Jew in America," people would get rightfully riled up. Yet we've allowed a single entity to assemble a database orders of magnitudes more detailed.

> If the government said "let's make a database of every Jew in America," people would get rightfully riled up.

This is an interesting case. You're completely correct, but it's basically a special privilege that Jews enjoy (and that has spread to a taboo on asking about religion on the census). If you're part of any other demographic group and you don't want to be counted, you get yelled at for not seeing the grand scientific project of the US census in the proper light.

But I don't really see that a database of every Jew in America is more inherently suggestive of abuse (if you're a Jew) than a database of every black in America is (if you're black).

> it's basically a special privilege that Jews enjoy

Any group with (a) a living memory of violent ostracisation and (b) some ability to conceal their group membership knows this.

Gay men and women, for example, would fight against a sexual orientation question on the census. (One wonders if Tim Cook, had he not been born a gay man in Mobile, Alabama, would be as sensitive to privacy issues.)

Side note: I used Jews in my example because we have a case study for them. In countries where databases of religious affiliation were kept, oftentimes for tax purposes, the Nazis took advantage.

Counterpoint: http://reason.com/blog/2018/04/06/i-dont-want-to-tell-the-fe...

> For the LGBT question, the exact opposite is happening: People who want a head count of gays and transgender people believe the data will then be valuable in influencing federal policies and spending on projects that benefit LGBT people—or, more accurately, to benefit certain LGBT organizations.

See also the effort by certain ethnic groups to split into a special MENA category on the census, despite having passed as white successfully for the last entirety of history.

It's not about ability to conceal group membership.

My opinions are influenced by (if memory serves) some Microsoft researcher who years back said something along the lines of how having all the data points is how computers can truly benefit an individual.

You can see the early fruition of this by looking at how useful Google Now/map planning etc are.

Computers could be our all knowing assistants.

That's on a very surface superficial level.

Imagine the kinds of benefits when this all knowing entity is applied to medical issues. Real substantive progress could be made.

However, today's mismanagement of data could scare people off all knowing entities for good & end up severely limiting real progress.

As with everything else, all that data is rife for abuse if the necessary legislation & technical limiting/auditing regarding access/use are not first put in to play.

There are tremendous benefits to centralizing these data. But that power cuts both ways. With the benefit of hindsight, Mark Zuckerberg and his acolytes created about the worst possible culture we could have picked, as a society, for entrusting these data to.

Worst possible? Pffft.

Reply on news.ycombinator.com