It seems like highlighting how anonymization is a lot harder than a lot of people assume is a really useful service. If researchers can do it, without any particular secret sauce, so can a lot of other people. (Unless I'm totally misunderstanding your comment.)

Some of Sweeney's most well-known work in this area is from the LATE 1990s. She was sounding the alarm about problems with anonymized data in medical datasets: https://en.wikipedia.org/wiki/Latanya_Sweeney#Medical_datase...

Her work almost certainly contributed highly to awareness of these risks.

More recently she has apparently worked on things like protecting voting rights in the US by notifying voters if their registration records change.

But, yeah, at some point in the 90s, Massachusetts decided to release some "anonymized" health records for research purposes (I think just state employees). One was governor William Weld who obviously had a lot of public information widely available. As I recall, Sweeney wrote the governor's office a bit later basically saying "I have your medical records."

I used this as a slide or two in some AI presentations in the mid-2000s or so pre-LLMs when I had some peripheral involvement with some of the privacy-preserving research going on (differential privacy, multiparty computation, fully homomorphic encryption). Haven't really followed most of this for a while.

https://en.wikipedia.org/wiki/AOL_search_log_release

You can be pretty sure some three-letter agency trash had been already using it around the world along with shady spyware startups.

Most other professions have you take ethics classes, have ethics boards and even ethics legislation. We're severely lacking in this area as a community. It really shows when every year there's a new company building the Maximum Oppression Orb from the book Dont Build the Maximum Oppression Orb. Its like we're dealing with the moral equivalent of a mentally challenged person all the time

The requirements for this sort of stuff come from top down. Do you expect C-Level and and the top layers of sycophants beneath them to be ethical?

If they hadn't done it, you can bet that bad guys would have done it instead (and maybe were already doing it). What the researchers did is publicly show that the existing schemes were broken, hence motivating the design of better schemes.

Like, you fundamentally misunderstand computer security research if you think that shitting on people publishing attacks is a good thing.

Should issue the award!

You're assuming Hollywood studios would ever release their content without DRM of some kind. They were quite content to ignore computers entirely if they didn't bend.

The world where Widevine doesn't exist isn't a DRM free one; but a world where an iPad or Smart TV can stream and a PC can't. I would support giving them an award though for "most repeated invention that keeps failing."

I don't see it anywhere.

I also find it hard to get offended about because there is basically no job, outside of tech, which doesn't involve physical location. >95% of jobs require physical location. Do you think a concrete worker, a plumber, an electrician, or literally anyone who works with their hands, has a right to location privacy? What does that even mean? "I'm totally clocking in to work today and totally installing a light fixture for a client right now and I won't tell you which one"? "I'm totally making a cappuccino for an old lady right now at one of our 30,000 branches, but trust me, you don't need to know which one"? Whining about this is extremely hard for me to generate sympathies for.

Overall it's just kind of a yucky and weird feature; when I worked in an office I really didn't really want my coworkers having a real-time automated feed about where I'm located and one of my chores as a manager was always picking a seating position where I could at least take the drive-by questions before my team got interrupted, which stuff like this bypasses. I could actually see it being useful for field-deployed employees but it's not part of the stated implementation and most people in that scenario already have a solution for that.

I agree that the typical HN-meltdown isn't warranted here; the HN Meltdown Factor on anything related to privacy, cryptography, and security lately has gotten really out of hand (the post you're replying to is a perfect example, actually). But I also don't think these counterpoints are very strong; they're justifying other useful features and products that almost everyone already has. It's weird to me that Microsoft haven't either clarified or backed down on this one given how much press it's gotten vs. the seemingly tiny advantage the feature presents.