29 Jan 2026 Search

After I put up the Google “Privacy Sandbox” timeline, it seems a common reaction was that people remember being discombobulated by the whole saga. Now that I have had a chance to think about it some more, maybe the problem with “Privacy Sandbox” wasn’t that it didn’t give the privacy people the kind of privacy they wanted, or that it didn’t give the ad people the ad features they wanted.

Maybe the problem was that at a deeper level, “Privacy Sandbox” was an exception to two of the useful heuristics we use to make sense of some complex business decisions.

“If it’s open source it must be honest”

Remember the original promise of open source, as laid out in the Open Sources Book from 1999? “Many eyes make all bugs shallow” or something? What a decade. Like the AI agent said, “The peak of your civilization.” (For what it’s worth, The End of History was supposed to be in the 90s, too.) In that optimistic time, one of the original ideas of corporate open source was that somehow software companies are going to be motivated to work in the user’s interest if the code is available for anyone to review.

And yes, open source worked for Linux and other high-profile projects with lots of interested developers, but for “Privacy Sandbox,” not so much. The project was spread over (FIXME: actually count these) n different GitHub repositories with m different owners, and discussed at a variety of online and in-person meetings. (Chaff doesn’t block radar, it returns extra radar signals so the real bomber is hard to pick out.) icymi: “The Secret History of Silicon Valley” Just sticking an open source license on something doesn’t make it honest, as long as Google can count on concealing it through complexity—and the people qualified to review Privacy Sandbox had more urgent issues on their to-do lists. Dealing with Google’s current ever-changing advertising shenanigans took priority over reviewing hypothetical future ones.

“This can’t be crime, because people have better opportunities”

In a legit economy, the kind of people who can get hired at Google (technically knowledgeable, work well with others, good communications skills) generally have much better prospects choosing honest work over crime. A good green flag for a high-trust society is expected returns to crime that are consistently lower than expected earnings from honest work. And until recently, that has mostly been true. At the low end, dealing crack pays less than minimum wage. At the high end, Bernard Madoff had a few hundred million dollars left when he was caught, after stealing about fifty billion. And in the middle, an employee of a legitimate IT company, on average, could expect lifetime earnings greater than those of a criminal hacker or online scammer.

But today, with the normalization of Scam Culture, the remuneration for honest and not-so-honest career choices works out a little differently. The Big Tech oligopoly extracts so much un-earned value from the legitimate economy that they push the expected returns from legit business down. Meanwhile, the companies that funded the “Epstein Ballroom” for the White House were not wasting their shareholders’ money—they are making a bet that crony capitalism beats the expected returns on legit investments.

The “Privacy Sandbox” project really only made sense in context. For example, at the same time that “Privacy Sandbox” was going on, Google was also changing the trademark policy for search advertising, to make it easier for one advertiser to use another’s trademarks without being detected. (The trademark owner must somehow find the infringing ad and go through a reporting process, which has limited effect since the ads are personalized and hard to find.) Today, Mac Malware is sneaking into Google ads not because of some wizardry by the malware hackers, but because of a deliberate decision by Google to expose search users to more infringing, often risky, ads. So the old heuristic that a qualified, effective person has better prospects in honest work no longer applies.

So whose lobbyists are they anyway?

While people with a direct experience of “Privacy Sandbox” are likely to abandon the heuristics that might have worked well before today’s Big Tech/crime situation, there are still a bunch of situations in which 90s optimism is not serving us well.

One big example is the ongoing debates over state privacy laws—where organizations that are supposed to represent news publishers are coming in on the side of Big Tech and against the interest of their own members. I wrote about that situation in a new column for the Reynolds Journalism Institute: Big Tech runs counter to journalism values. So why is the news industry helping tech take over?