As a small aside, The capitalist's answer is that regulating companies to prevent them from steering to the most profitable items is both impossible to be adequately done and prohibitively costly. Even assuming cost isn't an issue, it's hard to imagine such regulation to be equally applied to all market participants (or to be equally effective). So we would be left with companies that cooperate and others that defect, and the defectors would be favored (more profitable) and outcompete the cooperators in the long run.

So instead we start from the assumption that companies are greedy and let them compete to offer customers the best value -- and if that value comes (at least in part) from not being tracked, companies that do not track will attract more customers. We probably just haven't made enough of a fuss about it with our dollar-votes.

For what it's worth, I block all ads without giving it a single thought. The way I think about it is that on the flip side of the prisoner's dilemma, I'm just defecting like some companies would. It's a race to the bottom in terms of the trust between customers and companies, but I didn't make the rules of the game...

The assumption in this argument is that consumers are able to observe and quantify the harm of tracking more effectively than regulators could create laws against data collection.

Personally I think the success of either one comes down to cultural factors that are currently stacked in favor of advertisers.

Not necessarily, because creating laws isn't enough to regulate. You also need to enforce such regulation, and that's where the challenge lies. The argument assumes that in the long run consumers are more effective at rationalizing their choices than the government is able to appropriately enforce regulation.

Alternatively, it assumes the cumulative harm created by the disconnect between current customer behavior and rationalized customer behavior (i.e. prior to their rationalizing the status quo) is less than the cumulative harm caused by inefficient regulation, including the defector's problem mentioned earlier but also other negative externalities such as encouraging corruption / fraud (which itself requires further enforcement)

The way you are framing this serves only to reinforce talking points from those who are benefitting from the current situation. For instance, you're basically stating a priori that regulation is expensive and ineffective, and as evidence you talk about long tail of enforcement and defectors. But the ad revenue market is so consolidated you only need to enforce on a handful of players (Google and Facebook basically). The idea that defectors would then swoop in and create a massive enforcement problem is not substantiated. There have always been fly-by-night operations in all types of business, and they don't gain a huge advantage that catapults them to overnight success just because others play by the rules. No one is saying enforcement is easy, but to assume that it will be fatally flawed if it can't be perfectly applied to everyone plays right into the hands of those who are profiting from abuse of our data.

Now on the other side framing this as a "customer value" problem that will be sorted out by the hand of the market is just pure capitalist oligarch koolaid. How do you expect customers to have any sense of what data practices are behind their every day digital product choices, let alone quantify that into a dollar value? And even assuming they do all that, where are the market choices when everyone behaves this way? Even where there is theoretically a choice, many services have a huge network effect that makes a consumer's choice all but pre-ordained.

We need to have a reality check here. Markets are great when they work, but they are not magic and can not solve all problems.

In the real world market inefficiency creates local maximums, which can then be leveraged to implement policy. The most lucrative policies are to make those maximums even stickier. Advertising itself is a prime example of this - in a perfectly efficient market, once a brand became well known you'd think that additional money going to advertising would be a waste - causing the company to be less competitive and they'd dial it back. But instead what saturating advertising actually does is crowd out any new competitors that might come along. So as a customer, you're effectively overpaying so that you can have less choice!

This effect becomes even more relevant as the costs of production drop to zero, as an upstart competitor cannot get a leg up by optimizing production - in other words, the brand itself is a larger component of the "value". And on the larger topic, these days large corporations are declared "too big too fail" and bailed out by the central bank, rather than letting market mechanics assert themselves in even the most pressing cases.

Effective libertarianism involves recognizing that corporations and government are not dichotomous types of entities, but rather that both lay somewhere on a continuum of coercion. If the companies offering a product or service effectively move in lock step on some policy, then your main ability to reject that policy consists of going without that product or service. This is perhaps easier, but of the exact same vein, as needing to physically move to reject specific laws.

True, but only because that’s the wrong approach. The correct regulations are the ones that result in more competition. That’s treating the cause rather than the symptom.

If Amazon had to seriously worry about competitors, they wouldn’t be focused on selling overpriced garbage. Why? Because customers will notice that Amazon sells overpriced garbage, and will instead buy from somewhere else.

I don’t know what those regulations might look like, but I do know that pretty much every single “evil” behavior in the market can be solved by throwing in competition. It’s not always possible (e.g. maybe someone is locked in to a single vendor due to a bad contract), but when customers are given choices, the choice that offers the best value will survive in the long run.