A few thoughts on TikTok: I joined TikTok during the pandemic. I watched trends rise and fall. I absorbed the platform’s new vocabulary, born from censorship and irony. Suicide became “unalive.” Ass became “dat ahhh.” It was absurd but poetic in its own way. Language, like sea glass, reshaped itself to fit the medium—what was once sharp worn down into something vaguely familiar but entirely different.

These codes are legible if you spend enough time marinating in the app’s glossary of brain rot. But the more dangerous grammar is subtler. The algorithm doesn’t just reflect your preferences—it hones them, coaxing you toward extremity. As wsj.com/podcasts/tech-n… it, TikTok “learns your deepest desires” in an eerie few beats, feeding you more of what keeps you watching. Every observable input is tracked: your search queries, the velocity of your scrolling, and the pauses between swipes. Anything over five seconds counts as a view. Each becomes a signal, aggregated and weighted against billions of others to predict your next move. The result is not a mirror, but a statistically optimized version of you—built to keep you watching.

Over time, it can feel more intimate than self-knowledge—drawn from impulse, from the hidden parts of the mind that rise only when you are certain no one is watching. The algorithm pulls from the same place as your late-night searches: the weird medical question at 3 a.m., the kink you’ve never said out loud, the idle thought about disappearing or having an affair. Your search engine has always known. Now that information lives inside every feed you open, refined and fed back until it can predict not just your habits but the outline of your most private impulses.