People love mocking politicians when they reach for a folksy metaphor. When former Vice President Kamala Harris told podcast host Kurt Alexander, better known as Big Boy, that social media algorithms try to dumb people down by assuming you only want more of what you already saw, the internet did what it always does. Clips spread fast. Jokes followed immediately.
Look past the punchlines, though. The underlying warning about algorithms is not as silly as the critics made it sound. Recently making waves in this space: Litigating National Security Blacklists The Anthropic Precedent.
Harris showed up on Off Air with Big Boy in September 2026 and riffed on how feeds on platforms like TikTok and Instagram shape what people see. Her phrasing was clumsy. Her flower example got laughs. Yet if you spend any real time watching how modern recommendation engines trap human attention, you recognize the exact mechanism she was stumbling toward.
What Kamala Harris Actually Said About Flowers and Feeds
Around the twenty-four-minute mark of the conversation, Big Boy brought up AI, social media, and how feeds push tailored messages depending on who you are and where you look. Harris argued that platforms try to dumb people down by operating as if nobody has interests beyond what they have already viewed. Then came the line that launched a thousand memes: if she looks at a yellow flower, the algorithm acts like she has no further interest and only has room for yellow flowers, as though her brain could not handle the idea of a pink flower. Soon after, the chat drifted into collard greens and cooking. Further details into this topic are detailed by Ars Technica.
Critics jumped on two things right away. First, commentators pointed out that during the Biden administration, Harris had been tapped to lead AI efforts, making a casual, floral explanation an easy target for opponents. Second, plenty of listeners felt she talked in circles instead of giving a crisp technical breakdown.
Honestly, that criticism is fair on style and dead wrong on substance. Politicians rarely talk like machine learning engineers on a laid-back radio and video podcast. When someone says a feed acts as if your brain cannot handle a pink flower, they are not claiming a piece of code literally insults your IQ. They are describing a real frustration anyone who has ever lingered three seconds too long on a video knows well: you watch one clip about a niche topic, and suddenly your entire feed looks like a hall of mirrors built out of that single moment.
Why Recommender Systems Act Like Yellow Flower Traps
To see why Harris hit a real nerve, you have to look at what recommendation systems optimize for. They are not built to broaden your mind. They are built to predict what keeps your eyes on the screen.
Research published in Behavioral Sciences in October 2025 notes that social media platforms rely on complex recommender systems that filter, rank, and suggest posts using signals like clicks, shares, and watch time. Because those systems score content by predicted engagement, they naturally lean toward material that matches your past behavior. If you pause on a yellow flower, the model tests whether more yellow flowers keep you scrolling. Often, they do—at least in the short term.
Researchers draw a useful line between a filter bubble and an echo chamber. A filter bubble comes from algorithmic personalization narrowing what gets shown to you based on inferred preferences. An echo chamber grows out of how people group together socially and reinforce shared beliefs while pushing opposing views to the margins. In real life, both feed each other. When your feed keeps handing you the same narrow slice of topics, it becomes easier to mistake your current feed for the whole world.
A study of Generation Z social media users aged 18 to 27 published in the International Journal of Research and Innovation in Social Science in June 2026 found that perceived recommendation algorithms and algorithmic awareness accounted for 37.6 percent of the variation in filter bubbles. More interestingly, greater algorithmic awareness was linked to lower filter bubble perception and stronger critical information-seeking habits. Knowing how the machine nudges you changes how you look at what it serves.
There is also a classic tension in recommendation system design documented in computer science literature, including a systematic review on arXiv. Systems tuned tightly for prediction accuracy and immediate engagement often sacrifice novelty and serendipity. Push too much unfamiliar material, and users may scroll past or get bored. Keep serving safe, familiar matches, and engagement stays high—even if intellectual curiosity gets narrower by the week.
Call that "dumbing people down" if you want a blunt political phrase. In cognitive terms, if your daily information diet shrinks to repetitive hooks, outrage bait, and hyper-narrow niches, you get less practice weighing unfamiliar ideas or sitting with nuance. That is not a sci-fi plot. It is a side effect of chasing watch time.
Where the Dumbing Down Claim Goes Too Far
Still, if we are going to be honest with ourselves, Harris's framing has real blind spots.
For one thing, algorithms do not sit around judging human intelligence. They run statistical predictions over massive behavioral logs. They do not care whether a pink flower is deeper than a yellow one; they care which one predicts the next tap or next ten seconds of view time. Treating a mathematical ranking loop like a snobby gatekeeper makes the system sound more intentional than it is.
For another, blaming the algorithm for everything lets users off the hook too easily. You still choose what to search for, which accounts to follow, what to linger on, and what to share. Studies on user interaction and social network diversity show that a person's own habits and connections shape how uniform or varied their feed feels. If you only ever reward loud, one-sided, or superficial posts with your attention, the feed is answering your behavior as much as shaping it.
Political context matters here, too. Harris spoke during a period of intense media scrutiny and ongoing speculation around her political future—including whether she might run for the White House again in 2028. Some observers backed her call for independent journalism and sharper critical thinking around targeted online misinformation, while harsher critics argued that anyone who spent years in public office or shaped tech policy should speak with more precision about how recommendation models work. Both reactions show how quickly a simple conversation about feeds turns into a test of political credibility.
And let's be clear about what the evidence does and does not prove. Researchers broadly agree that algorithmic curation can narrow exposure and amplify polarization under certain conditions, yet findings vary across platforms, user groups, and measurement methods. Saying algorithms "dumb people down" works as a catchy warning on a podcast. As a scientific claim, it is a shorthand label for a messier mix of business incentives, design trade-offs, and human habit.
How People Actually Psych Out Their Feeds
In her interview, Harris talked about figuring out how to "psych it out a bit" so the feed does not box you in. Having spent years watching how people try to outsmart their feeds—and seeing how often clever hacks fizzle—I can tell you that fighting an algorithm head-on usually fails.
People try all sorts of tricks. They mute keywords one by one. They hit "not interested" on every stray post. They create burner accounts to reset their slate. Those moves can help a little, yet they rarely last if your everyday scrolling habits stay the same. A single accidental pause on a dramatic clip can pull you right back into the same lane.
What actually works is quieter and less glamorous. Start by changing where your curiosity begins instead of letting the "For You" or discovery stream pick your menu every morning. Open a search bar with a real question in mind. Read full articles from outlets or writers with different angles rather than relying on twenty-second summaries. Follow people who think differently than you do without turning every reply into a fight, which research on social network variety suggests can help break up overly uniform feeds.
Pay attention to how you react, too. Notice what makes you tap out of sheer boredom versus what makes you stop because you are annoyed. Anger is one of the cheapest fuels on the internet; if you feed it, the system will gladly serve you more of it all afternoon.
You do not need a grand theory of artificial intelligence to spot a narrow feed. Just ask yourself a simple question the next time you close an app: did I just learn something real, or did I just get shown ten versions of the same yellow flower? If the answer is yellow flowers all the way down, log off, go outside, and go find a pink one yourself.