This time last year, U.S. job platform Handshake was a modest startup generating just $5 million in annual revenue. This year, it has surged to $1 billion in annual revenue—all thanks to the immense profits of being a “human traitor.” By “human traitor,” I refer collectively to data labeling companies that act as guides for AI, connecting highly educated or specialized talent to model developers to solve tasks, thereby enhancing AI’s domain-specific capabilities and accelerating human replacement. Handshake had spent 12 years in the niche college recruitment market, barely scraping by. Last year, it had a revelation: since it already had access to the entire U.S. university community, why bother helping students laboriously find jobs when it could sell them directly to wealthy and generous AI giants? Within a year, Handshake’s revenue skyrocketed 200-fold. After deducting payments to the task workers, it still pocketed $300 million in net profit, making Handshake the fourth data labeling company in recent times to join the $1 billion annual revenue club. Isn’t selling shovels in reverse also a massive business?