Essays

For a century the advertising trade repeated a line about wasted spending that almost certainly nobody ever said. A run of field experiments — a search-engine blackout at eBay, a statistical reckoning from twenty-five randomized trials, and a method for testing by geography instead of by person — has finally started to answer the question the myth never really posed.

Ad fraud is usually pitched to marketers as a security problem — something to patch and move past. The practitioners who actually keep their numbers trustworthy treat it instead as an ongoing discipline of data hygiene, because every model built downstream inherits whatever contamination gets let through.

Four field studies — a call center, a writers’ desk, a consulting floor, and a software team — converge on the same finding: raw access to AI barely predicts who gets ahead. What predicts it is whether someone built the habit of checking the machine’s work, and learned exactly where to stop trusting it.

The same sixty minutes sells for $21 if you are an average freelancer and for nearly $3,000 if you are a senior partner at an elite law firm. A history of the billable hour, and the decades-old rebellion inside accounting and consulting trying to kill it in favor of pricing outcomes instead.

Most new PhDs used to become professors. Most no longer do, and the numbers on where they land now say less about individual failure than about what doctoral training actually builds. A look at what travels from the lab into industry — the method, the writing, the discipline of peer review — and what quietly stays behind.

Britain’s Global Talent visa admits leaders and future leaders in science, technology and the arts — but the decision that matters happens before any immigration officer opens a passport, when an outside panel reads a CV, letters of recommendation and a stack of scanned documents. What counts as proof of recognition turns out to depend heavily on which desk is doing the reading.

Across the OECD, people who have just crossed a border are more likely to start a business than people who never left. The honest explanation is less a story about grit than about closed labour markets, informal credit networks, and migration itself acting as a filter for risk-takers.

A wave of research on the “protégé effect” suggests that explaining what you know to someone else is not time taken from mastering it — it may be one of the more reliable paths to mastery itself. From a nineteenth-century surgical ward to a psychology lab in St. Louis, the evidence keeps pointing the same direction.