The Oracles in Our Feeds

While scrolling TikTok on a hot summer day, I stumbled on a video engineered to stop my thumb. “This video has no tags. If it found you, it was meant for you,” a tarot reader said, a jewel-bright deck fanned between her hands as a set of waiting fates. Though I knew no spirit had actually summoned the video to my screen, I found it hard to dismiss it as a coincidence.

The TikTok reader turned over a card, then another. She narrated a scenario: someone from my past was thinking of times we had spent together; thorny conversations I hadn’t yet fully finished would return; a decision I had been avoiding would soon demand an answer. Her predictions left enough space for me to fill their ambiguities with names and memories from my past, and with my own internal anxieties.

Since then, my feed has delivered dozens of similar prophecies. Each announces, with a certain gravitas, that it matters. You were meant to see this! This message found you at just the right time. Whatever you do, do NOT scroll. Someone in your life whose name begins with the letter “H” is plotting your downfall (I sent this to my sister, chuckling, with a “??”). This video has no hashtags, so it was truly meant to find you today!

For most of human history, uncertainty has attracted interpreters. Ancient Greeks consulted oracles, watched birds fly, examined dreams, and read sacrificed animals for signs of what awaited them. The oracle promised that the world’s chaos could still contain messages if one knew how to read them. Ordinary events were interpreted as signs of something in our foreseeable futures. A bird crossing the sky was not simply being a bird; a dream was not simply a dream.

Our recommendation systems belong to an entirely different order of knowledge. Their predictions emerge from bits and pieces of data collected from us and behavioral inference, not the supernatural. They gather small traces of our behavior—which jokes make us laugh, which strangers catch our attention, and which cultural moments or trending topics have us hooked—and use these patterns to decide what to show us next.

Spotify tells us what music era we will enter next; TikTok determines which corner of culture will appear when we swipe up; dating apps, even, reduce an impossibly large group of people to a handful of strangers presented as conceivable future partners. This question of what awaits you in the future, one that has captured our imagination and anxieties for centuries, has not disappeared. In many ways, this language of prophecy has become a language of prediction.

These algorithms increasingly help organize our futures. When TikTok notices us lingering over one particular video, it sends us three more, and soon it develops an accidental curiosity in us that may become an interest, a passion, even. What begins as a guess by TikTok about who I am can help create who I become. A familiar joke we see appearing on our social media feeds nowadays is that our algorithm “knows us better than we know ourselves” and sayings along the lines of “not my feed calling me out!” Even though these recommendation systems don’t know us in any human sense, the judgments they make from traces and patterns in our digital behavior can feel uncannily intimate on the other side of the glass.

A video names a fear we haven’t yet articulated, and a very specific song might arrive the night we need it. A prediction, whether from a tarot reader or a stranger who happens to share your experience, can land close enough to the bone that the algorithm’s calculations begin to make us feel deeply recognized and seen in our humanity. This recognition can oddly begin to resemble what we think is perhaps fate. And even understanding this machinery doesn’t necessarily break the spell, because any rational explanation we can devise for why something appears on our feeds can still sit comfortably beside this colder shiver in our minds: what if?

That may be the peculiar enchantment of the algorithmic age. Our reason has not entirely displaced superstition. We’ve built great, incredibly sophisticated systems of computation, only to discover their predictions can awaken impulses in humans far older than computation. The desire for signs, the fear of fate, the hope that the universe (or at least the feed) has noticed us. The crystal balls of the witches in our childhood movies have taken a new shape: now just a black mirror held intimately in our palms, glowing with a stream selected just for us.

That is why I still stop when these videos appear, and why my thumb still hesitates to skip any similar video that pops up on my feed. I know the cards don’t really know the real me, and the algorithm doesn’t care whether any prophecy comes true. But for a few seconds, its selection can still feel uncomfortably close to reality. Beneath the reader’s card, another system is already studying whether I stay.

Scarier still is the possibility that viewers who feel recognized, feeling a part of themselves seen—and that this recognition might, to a certain extent, replace the capacity for real human connection that comes from in-person encounters. This also shows up in our growing dependence on conversing with artificial intelligence systems. A 2025 study by Common Sense Media found that roughly one-third of teenagers who had used AI companions had discussed serious matters with them instead of real people. Some have even turned to chatbots for therapeutic purposes, even though Stanford researchers warn that such systems can respond dangerously to mental-health concerns. The danger, perhaps, is that feeling understood by machines may gradually replace the experience of being understood by another person through actual human connection.

As algorithmic systems grow more capable and fluent at predicting human behavior, another question becomes harder to dismiss: at what point does prediction stop describing the future and start quietly arranging it, becoming a blueprint for who it wants us to become? If a machine learns from us while continuously deciding what we encounter next, who, exactly, is reading whom? Perhaps the greatest power of these oracles, assigned to us by our feeds, is not simply to prophesy and predict our futures, but to eventually make their own prophecies come true as they shape who we are.

Mujtaba Raja ’30 (mujtabaraja@college.harvard.edu) is comping the “Harvard Independent.” He would stop doom-scrolling, but his algorithm says something important is coming.