Sure, I don't expect certainty, just some justification. Are you saying it's just wild speculation on your part?

> You are the one who is making extraordinary claims based on extremely thin evidence. You are claiming with great certainty that giving up any private data to advertisers makes democracy fall apart,

First of all, I don't actually make that claim. I can see how you maybe can get the impression from the short version (as I mentioned, it's kinda hard to condense a complex topic into an HN comment), but really, it's a matter of the severity of the possible consequences combined with non-trivial probability. Which is one reason why the comparison to laws regulating how you cannot prove to other people how you voted in an election is an insightful one: Those restrictions don't exist because the probability is particularly high that any given election would be corrupted, they exist because the downside is massive if it happens even once under the wrong circumstances. We make every single election more complicated than usually necessary, in order to help protect against the few cases where the consequences would be catastrophic, even though there is not even remotely a guarantee that that would ever actually happen.

An important question to ask is: Does a stable democracy protect against those risks, or is preventing those risk by other means what keeps the democracy stable?

> and yet you are unable to point to a single stable democracy that has fallen apart for that reason. Your sample size is zero.

May I suggest that your categorization is counterproductive?

Suppose I build a bridge from papier-mâché made from cardboard that was used to package the iphone 7. You might argue that that's not a reliable bridge and it shouldn't be opened to traffic because of the risk that it might collapse, given the track record of papier-mâché and our knowledge of physics. Now, I would reject all your arguments by pointing out that no bridge built from cardboard that was used to package the iphone 7 had ever collapsed and that your sample size is zero, and that we therefore have no clue whether that bridge is a risk to public safety. Would you agree?

It's hardly surprising that that specifically hasn't happened yet, given that the technical developments are relatively new, and some of it is yet to happen. That doesn't mean that we can't draw any conclusions from similarities to other power structures in history, or that we have no clue how current or expected future technology could be abused.

> This is a very common way in which predictions of doom tend to go wrong. Take one aspect, follow it to extreme conclusions as if there were no other variables or counter forces or reactions to what is happening. This isn't logic or consistency.

It seems like you are just filling in the holes in what I wrote with the most idiotic ideas that I could possibly have, and then you pretend that strawman is what I actually wrote.

BTW, as far as I can tell you are making that exact mistake in the opposite direction: You find some aspect that isn't necessarily quite as bad as my simplified model desribes it, as there is some other influencing factor or some counteracting force that could decrease the impact, and from that you conclude that therefore a catastrophe is completely unrealistic.

> And this lack of a good model starts where we haven't even discussed what privacy or private data is in the first place. It's definitely not one thing that you can switch either on or off. We always have to trust someone with some of our data. Much of our "private" data isn't ours exclusively as it may be about interactions with others.

As you say, we haven't discussed it. From that it doesn't follow that my concepts are idiotic.

And yes, I agree that privacy is not a binary distinction, nor is privacy an end in itself, nor is some sort of "total privacy" a sensible goal, let alone achievable.

The root problem is that personally identifiable information confers power. Power is not inherently bad either, nor something that could realistically be avoided completely. But if there is one lesson we as humanity should have learned from history, very much the hard way, it is that concentrations of power are dangerous. There is no clear line between how much concentration of power is harmless and where the danger begins, but it still is quite clear that too much of it in one place/in one person's hands is a terrible idea.

If you think about it, the whole point of democracy (and the rule of law and checks and balances) is to prevent concentrations of power. And we accept a lot of inefficiency in the government, compared to dictatorships, in order to maintain that distribution of power. The laws governing the election process are only a small part of how we build democratic societies with lots and lots of expensive safeguards to prevent the concentration of power.

> The data we give up to advertisers isn't used exclusivly for advertising either. It's also used to provide some of the services we want (like search or spam filtering).

I think most of the data shouldn't really be necessary to achieve a sufficiently similar result.

> "Ban it all!" is neither proportionate nor even possible.

Depends on what exactly you mean by "it all".

> And even if I did follow your argument that giving up any privacy at all makes democracy fall apart, I see no alternatives to advertising that are better for privacy.

Is it intentional that you are equating advertisement and surveillance?

> Just look at the numbers involved and you will quickly find that funding everything like Tor or a few IRC channels is impossible. Donations, self-funding and sponsorship involve payment and are subject to extremely detailed surveillance. There is nothing more heavily regulated and traced than payment and you can be absolutely sure that governments will never allow large anonymous money flows because everything that could fund something like a Google data center could also fund organized crime and terrorism.

Why is it necessarily a big problem that payment is subject to surveillance in this context? How is surveillance of the fact that the average human in the industrialized world pays a certain amount of money per month for telecommunication and data storage services equivalent to surveillance of the contents of their interactions with each other?

Also, who says that we need something like a Google data center?

> That said, I also see a lot of negatives in the excessive and brazen behavior of the ad industry, and I also think there is a need for regulation. A better regulated ad model is our best shot at more privacy and fewer negative externalities.

And I don't think there is one that's going to improve much other than to forbid the collection of personally identifiable information, possibly unless it's completely optional (not in exchange for anything) with requirement for informed consent before any data is collected and with the option to revoke consent at any point, in which case all collected data about a person would have to be deleted immediately.