End result: Signal is a real, usable tool, the recommendation of which is fully actionable security and privacy advice you can share with anyone who is capable of using WhatsApp.

Here's some HN discussion on the post, and a piece with counter arguments. The latter, I feel, misses the entire point of how urgent the need for usable privacy and security tools is.

https://news.ycombinator.com/item?id=11668912 https://gultsch.de/objection.html

That being said, a big part of me wonders if (assuming you could find lawmakers with enough technical literacy to propose it, which would be a pretty huge ask!) you might be able to get the best of both worlds by requiring sites like facebook to expose an API for programmatic access. That would solve the stangancy problems that Moxie highlights (not to mention sidestepping the need to get everyone to agree on any one protocol), whilst simultaneously forcing a whole lot of things to open up.

Alternatively: part of the reason these walled gardens exist is that they control the concept of identity. Not necessarily in the sense of "this is my profile", but more in the sense of authentication and authorization. If you moved that somewhere else -- say, in the browser -- it would make it much easier to build systems that are interoperable.

The problem with all newer systems is that they are identity based, which doesn’t scale down to all of the device/station use cases that phones exist for.

But if you start looking at those '12 users', you should notice that that the need is extreme.

This includes drug people and at-risk populations like sex workers: I'm not happy until drugs and sex work are decriminalized globally, so I want to protect sellers and buyers, a partial reason why I participate as a Tor relay admin. Then there are journalists and politicians. Apps like Signal, Wickr, Wire and Threema aren't as much the ecosystem we want as the silos we deserve.

That said, you'd be surprised how easy a sell Signal is in some cases.

I've gotten a bunch of friends over to Signal, from the likewise end-to-end encrypted Whatsapp. And after a family crisis, where a friend of the family died, I explained how the metadata of a mostly dormant Whatsapp group lighting up suddenly could be interpreted.

Facebook knows we're a family of aging parents and grown-up, mostly independent 25+ kids, so there might be a crisis or emotions whenever there are big bursts of activity. A crisis might make people more susceptible to predatory ads.

This was enough to convince my folks. So, family chats are now on Signal. The app is good enough. Granted, my parents and siblings are active in professions related to either legal or social services, so they understand the need for privacy and my overall desire to curb stomp the online tracking industry.

Sounds terrible? Good. That's what Facebook is doing now.

1. There's no effective signing on phone calls, so spoofing a phone number is easy. This makes it impossible to prosecute spammers.

2. There's no way to filter a phone call based on the content of the phone call before it arrives. It's possible to spoof senders on email and text messages as well, but it's also possible to filter these based on content before they reach your inbox. This isn't perfect, but it greatly reduces spam in text-based mediums.

You can eliminate ~95%+ of junk calls from the telco side.

The incentives aren't there, of course.

But even if not, it would already be a drastic improvement on what we have.

That is literally what Facebook had, and what people excoriate them over thanks to Cambridge Analytica. The Facebook haters are driving social network data into a silo; how else are you going to "keep control of your data"?

You can have privacy (or at least the illusion of privacy), or you can have access. You cannot have both.

I want to reach my friend Pete through "media". But I don't really care about his ad-relevant personality matrix, I just want the endpoint. The profile metadata is the aspect created and mistreated by FB, and abused by CA. I suppose scraping the proposed hypothetical API would be akin to getting a large list of valid emails, but it's not as useful for targetted agitprop.

You misunderstand what happened. 1) Users installed a quiz app, 2) Users explicitly accepted a permission dialog that included "information about your friends", 3) The quiz app saved this data, 4) Cambridge Analytica bought the app (and data).

If CA got your data, it's because one of your friends sanctioned the critical link in this chain.

Facebook at least provided this level of security, however illusory it may have been. In a decentralized system, anyone who touches a piece of data can pass it on to whoever they want. There is no way to enforce "my friends can look at my data but it's my data and they aren't allowed to pass it on".

Privacy advocates and decentralization advocates are mortal enemies; their only common cause is that it's popular to hate on Facebook these days.

And CA “leak” was way more tame than this, just the information displayed on your profile for all of your friends.

I can see arguments both ways. Some folks would be ok with it (they're your friends) and others wouldn't (you can't change the privacy setting on data that's already been downloaded).

Similarly with friends list. A person's list of friends is probably "their" data, even if it has your name on it. Would you be ok with them downloading this information?

The trade-off between control of some users and the privacy of other users is a complex issue. Let's not pretend there are simple answers to this.

The implicit assumption here is that the only people who can control your data are big, centralized organizations. Instead of big companies making little people subject to DRM, what if we turned it on its head and gave control to the people? What if we made big companies subject to Digital Human Rights Management? Basically, we legislate that social media, tech companies, and government organizations use Trusted Execution engines to handle private/personal data, and the user gets the keys and gets to revoke access from those organizations.

Let's say I put all of my friends' birthdays in my calendar. Now the data has left the silo and I can reshare it with whoever I want. Just how is anyone going to revoke that ability?

Also, it's not meant to be a complete solution by itself. It's meant to be a tool supported by other legislation. So it would totally be possible for Facecorp to crawl the web and put together a non-revokable database of private individual's information, outside of Trusted Execution engines. But to do so, they have to go to that effort, instead of just sweeping user data.

Having to ask the friends for approval doesn't scale.

There is the fundamental conflict between a social network platform and privacy.

IMO that would do a lot to reduce the monopoly power of social networks and platforms like Youtube, and be fundamentally fair.

There are third party apps that allow you to post to Facebook and check your Facebook status. Couldn't that other service post to Facebook and Mastodon on your behalf and read your friends profiles? I can't think of a reason it isn't technically possible for something like that to be created, but I haven't seen it created yet.

[0] https://www.w3.org/TR/activitypub/

If anyone deserves that sort of credit, it would arguably be Evan Prodromou and the original identica a decade ago. Though I could wrong, and unaware of anyone prior to him working on this sort of federated and open twitter clone.

And then things will get way, way worse than facebook, because nobody is there to censor anyone, and then the nytimes will be asking to censor the entire internet because reasons. No, please. Let people stick to facebook where they can be herded appropriately.