3 ms·
I understand it, I just don't find it compelling. Second order effect arguments are often weak, in this case - not using a cleverly designed system that doesn'
by fossuser 5y ago
I understand it, I just don't find it compelling.
Second order effect arguments are often weak, in this case - not using a cleverly designed system that doesn't track them because it could make some theoretical person more comfortable with the idea of tracking in general doesn't hold up to me.
The point of this is that it doesn't track them, but still achieves the goal of exposure notifications.
It's like the seat-belts are worse argument - if people wear seatbelts they'll drive more dangerously therefore seatbelts are bad. These type of second order arguments can sound contrarian or smart, but they rarely hold up to scrutiny. (Similarly masks are bad because they make people touch their face or w/e nonsense was pushed early on in the pandemic).
The average person is tracked completely by modern web companies, telcos, and random apps and they give up that willingly already - how does privacy preserving contact tracing make that worse? I don't buy it.
It's more likely using services that preserve privacy by design get people to understand why they're different than everything else, and why that's valuable. When you dismiss them along with the bad ones you make it harder for people to make that distinction.