- AI Vibe Daily
- Posts
- Can AI Predict Your Taste? Spoiler: Yes, and it is both impressive and a little creepy
Can AI Predict Your Taste? Spoiler: Yes, and it is both impressive and a little creepy
We like to think our tastes are deeply personal, but AI is getting better at reading them. If machines can predict what we like, does that make us more predictable than we want to admit?
Theo T
August 29, 2025
🔍 The Big Idea
Edmar Ferreira at Every decided to test whether AI could guess his personal tastes. He gave it a simple description of his preferences and asked the model to predict which articles he would upvote on Hacker News. The results were surprising.
The AI got it right more than 80 percent of the time, which is far higher than random guessing. In other words, with just a few hints, the model could map out his digital fingerprint. That feels like magic, until you realize it is really just math and pattern recognition.
But when Ferreira pushed further, asking the AI to predict which dense philosophy essays he would actually finish reading, the model failed. Taste is not just about topics. It is about mood, curiosity, energy level, and the undefinable spark that makes you think, “Yes, I want this right now.” AI cannot yet capture that part.
🧩 Where AI Wins and Where It Fails
AI Strengths | AI Weaknesses |
|---|---|
Predicting likes based on categories | Struggles with context such as mood or timing |
Spotting broad preferences quickly | Misses nuance and subtle reasons for choices |
Scaling predictions across thousands | Cannot always explain the “why” behind a choice |
AI is excellent at labeling you as the “type” who enjoys certain things. It can build playlists, recommend shows, and suggest articles you are likely to click. But the deeper essence of taste, the reason a specific song hits differently on a rainy day or why a weird poem sticks with you, is still yours alone.
💡 Why It Matters for Real Life
Your recommendations may feel eerily accurate. Netflix, Spotify, or Kindle might know what you want before you do. That is useful, but it can also feel invasive.
Discovery can shrink. If algorithms only feed you what they know you like, you may miss the thrill of stumbling onto something outside your comfort zone.
Your taste becomes a signal of identity. The less AI can predict, the more valuable your weird, personal choices become. Choosing something unexpected becomes your superpower.
Advertising shifts. Marketers who once chased demographics now target you through predictive taste. This changes how products are sold and who gets noticed.
💪 Try This Today
Test the limits of AI’s taste predictions:
• Write a paragraph describing some of your unusual favorites. It could be a niche podcast, a cult film, or a random snack you love.
• Ask an AI tool to suggest something new based on that description.
• Notice if it plays it safe or surprises you. Was the suggestion close enough to be useful, or too generic to matter?
• Finally, try the opposite: choose something random that you would never normally pick. Reflect on how much of your taste is predictable and how much is pure chaos.
🧭 Bottom Line
AI is already good at mapping broad preferences and predicting what you are likely to enjoy. But it still cannot fully capture the messy, situational, and emotional parts of human taste. That gap is your individuality, and it is worth protecting.
Want more content like this? Subscribe to our daily AI newsletter at AIVibeDaily.com