On Monday I took my dad to a medical appointment, then walked his dog through the park near his house while he recovered on the couch. Midway through the walk I spotted a Little Free Library box, opened it, and found Good to Great by Jim Collins — a book I'd been meaning to read for years but never gotten around to. I took it. The next morning, a Tim Ferriss newsletter hit my inbox announcing his latest podcast episode: an interview with Jim Collins.

My first instinct was to laugh at the odds of that happening, then half-jokingly say to myself that there isn't a tracking pixel or digital cookie in the world advanced enough to have made that connection. An open physical space, a borrowed dog, a weathered paperback in a wooden box. Surely that's outside the marketing tech stack.

Turns out, it isn't. Not anymore. I looked into it, and here's how it could happen.

My phone was in my pocket the entire walk. That's the entry point. Data brokers collect precise movement data from hundreds of millions of people primarily through third-party app SDKs, weather apps, navigation apps, coupon apps, that share location access well beyond their stated purpose, with companies processing upward of 17 billion location signals from around a billion mobile devices daily.¹ The park, the route, the duration: all logged, all available for purchase.

The mechanism most people don't know about is real-time bidding. When ad space is auctioned, advertisers are sometimes provided granular location and behavioral data about potential targets, and data brokers retain that information even when they don't win the bid.² The surveillance runs regardless of whether an ad ever reaches you.

The drive to my dad's place adds another layer. GM was found to have monitored and sold customers' precise geolocation and driver behavior data, sometimes as often as every three seconds.³ Not a government program. A car brand.

So the system knew I was in that park. It knew my search history, my purchase history, the articles I'd lingered on, the newsletters I was subscribed to, the interest graph assembled from years of digital behavior. What it couldn't see was me, alone, opening that box and reading the book titles.

That act of noticing, the moment of actual human attention, is the gap that made the whole thing feel like mysterious coincidence. That gap where no one but me could have known is why it still feels like magic. But that gap is also closing.

Consumer wearables with ambient object recognition, devices like Meta's Ray-Ban glasses, are already capable of identifying what's in your field of view and cross-referencing it against your interest graph in real time.⁴ If I'd been wearing a pair when I opened that library box, the system would have recognized the book cover before I finished reading the title. Everything that followed would have been less synchronicity, more algorithm.

The physical world is well-instrumented now; the last opaque variable is what you choose to notice. The Jim Collins coincidence felt like the universe wanting me to pay attention to something specific, putting it in front of me at every turn until I couldn't not see it, read it, listen to it. But soon, if not already, it may be data manufacturing coincidence, not just the universe. The data all around us provides invisible, granular information about who we are, but the one thing it couldn't track was the decision itself. The moment I reached in and chose to pick up that book. That's a surprisingly small space for magic to live in, and it's getting smaller.