How is it at all possible to certify that data has been deleted and no copies were taken..?
This is usually only noteworthy when it fails, but by and large it works. The friction involved if we did not generally accept someone's (signed, notarized, appropriately formalized) word as bond would cause our world to grind to a halt.
Even the flimsiest of technological improvements could radically improve the trustworthiness of the notarizations, but by and large the system works, so there's not much of a reason to fix it yet.
Including many technical spheres, and particularly including the vast majority of the software world; formal verification exists in software, of course, but is most notable in not being used for most of it. Yet, even without that, there's usually a release process that includes the moral equivalent of a certification—without formal verification or certain knowledge—that the software does what it is supposed to.
'Fake news' takes advantage of something much more simple. People believe what they want to believe. In other words the people led on by fake news tend to already want to believe what the news is about. When we see something that confirms our biases, many of us don't question it. And this is made even worse since people tend to work as apologists for fake news from sources they like, so instead of being able to discuss the issue of fake news - it ends up dividing people into splits on the 'fake' topic at hand, which in contemporary times is most often political.
Well, yeah, that's what I was referring to. If you perceive someone as being from your tribe then you are predisposed to believe what they say, and if you don't then you aren't. Nowadays tribes are much bigger than they used to be, and tend to center around ideologies more than geography. But otherwise it's monkeysphere dynamics at work.
That said, I do agree with your sentiment that the blockchain-as-a-solution space is getting a bit ridiculous. I can't imagine a way to translate this problem into a smart-contract-solvable form.
You can do some clever things to figure out who it is that leaks stuff, but that's another issue.
By, e.g., signing a paper which says that.
You probably meant how can you know that, which is a different issue than how can you certify it, and which amounts to a question of degree of internal accountability and control of data access.
Certification is about assuming responsibility for the truth of the thing certified, which doesn't actually require knowing it (though obviously knowing it to a reasonable degree of certainty makes it more comfortable to certify it.)
Seems to be plain copyright infringement... FB licensed the data to Kogan, not CA, hence CA violates FB's copyright on the data.
All our data associated with FB unique id is out there. My understanding this has stopped in 2015 but for sure?
Hopefully the data that is out there becomes less relevant over time.
[1] https://medium.com/@jamesallworth/what-the-f-was-facebook-th...
If someone leaks facebook data to a journalist, can FB demand that they delete it as well?
It costs Facebook nothing to ask developers to sign a contract written on virtual paper, yet can be used as a defense to cover their butt further down the line.
[1] http://www.businessinsider.com/well-these-new-zuckerberg-ims...
There's tons of substance to be discussed here without reaching for the global hash table (https://news.ycombinator.com/item?id=9722096). Fortunately the community has mostly been doing that, but let's not spoil it.
The point isn't to defend Zuckerberg, btw; it's to defend the quality of this community. If we're to realize our dreams of doing marginally better than internet median, we need people to check their worst instincts here.