Judge Richard Posner wrote about a huge variety of topics, including privacy. The Right of Privacy is a lecture from 1978. Since then, a lot less has changed than you would expect, considering the total number of technical and business transformations proclaimed over the years. Two of Posner’s assumptions, though, really have been affected by changes that happened between 1978 and today.

Transaction costs of selective sales of personal data. Posner gives an early example of cross-context tracking, “whether the law should allow a magazine to sell its subscriber list without obtaining the subscribers’ consent,” and suggests,

In the magazine case the costs of obtaining subscriber approval would be high relative to the value of the list. If, therefore, we believe that these lists are generally worth more to the purchasers than being shielded from possible unwanted solicitations is worth to the subscribers, we should assign the property right to the magazine, and the law does this.

And in a footnote,

A few magazines offer the subscriber the option of having his name removed from the list of subscribers that is sold to other magazines. But this solution is unsatisfactory to the subscribers (presumably the vast majority) who are not averse to all magazine solicitations.

Today, the costs of enabling people to manage use of their information are much lower. I don’t know if Judge Posner ever tried Rewarded Interest, but he might have appreciated the market design. It is possible to implement a property right in cross-context addressability. Related developments include MyTerms, a new standard for a web client and server to agree on a contract that specifies how personal data is used.

Both of those are still early, but they do show that Posner’s concerns about transaction costs and limited options are being worked on. (The path to adoption will need to include an adapter layer for the complex “consent” flows that annoy people today. Being nagged for “consent” every time is not going to be with us forever. It’s a step on the path toward a system that’s both accurate and convenient.)

Harms of cross-context addressability. In 1978, the problems of cross-context tracking were the problems of getting too much direct mail and too many phone calls.

Yes, some people back then tried to get as much direct mail as possible because they made it into artificial logs for their wood stoves, which is as just as much of a terrible idea for your lungs as it sounds like. In general, though, direct mail was a hassle. But, as Posner points out, it was arguably worth more to the sender than it cost the recipient. And, because postal mail has the United States Postal Inspection Service keeping an eye on it, a direct mail offer of 1978 would have been less likely to be fraudulent than an Internet offer of today. Posner assumed that cross-context data exchanges were between one honest company and another, which would have been a safe assumption to make in the direct mail environment of 1978.

Today, though, cross-context tracking is a much riskier proposition Not only is there no Internet version of the Postal Inspection Service, platform companies such as Meta and Google operate not just as perpetrators of direct harms to users but also as facilitators of fraud by others. And personal data passed to Meta or Google helps to train their machine learning systems for adversarial personalization. The Santa Clara County case should provide a better view into how even though Meta does operate an anti-fraud operation, it’s intentionally limited by the company’s pro-fraud design. And of course the Google search ads are bad enough to get an FBI warning. Losing your life savings to a scam targeted by Meta or Google is a lot worse than having to take out a trash can full of extra direct mail. I doubt anyone could have expected that back in 1978.