How to Use AI to Pick What to Watch, Read, and Listen To

You know the feeling. You sit down after a long day, open Netflix, and spend 35 minutes scrolling through thumbnails before giving up and rewatching The Office again. Or you finish a book you loved and have no idea what to read next. Or your Spotify Discover Weekly has been mid for months.

Streaming algorithms are supposed to solve this problem. They don't. They recommend things that are popular, things that keep you on the platform, and things that are similar to what you just watched. That's not the same as recommending something you'll actually love.

AI chatbots are shockingly good at this. Better than any algorithm I've used. And I'm going to show you exactly how to get great recommendations in about 30 seconds.

Why AI Beats the Algorithm

Netflix knows what you clicked on. That's about it. It doesn't know that you loved Severance because of the creepy corporate dystopia vibe, not because it was a "thriller." It doesn't know you're in the mood for something light tonight, not another heavy drama. It doesn't know you liked that one obscure Korean movie but hated the other three it recommended after.

When you talk to an AI, you can explain what you actually like and why. You can describe a vibe. You can say "I want something like the first season of True Detective but set in space" and it'll actually understand what you mean.

That's the difference. Algorithms match metadata. AI matches meaning.

Movie and TV Recommendations

Here's the prompt I use most often. Copy it, tweak it, make it yours:

"I just finished watching [show/movie] and loved it. What I specifically liked about it was [be specific]. I'm in the mood for something [vibe]. I've already seen [list a few obvious ones so it doesn't waste your time]. Give me 5 recommendations with a one-sentence explanation of why I'd like each one."

The key is being specific about why you liked something. Don't just say "I liked Breaking Bad." Say "I liked Breaking Bad because of the slow transformation of a normal person into someone unrecognizable, and because every episode felt like it mattered."

That specificity changes everything. Instead of getting a generic list of "shows like Breaking Bad" that you've already seen, you'll get picks that match the actual feeling you're chasing.

Here's a real example I used last week:

"I loved The Bear because of the intensity and the way it shows someone pouring their entire soul into their craft. I also loved how chaotic and stressful it felt. I'm not looking for another cooking show. I want something with that same obsessive energy about any topic. What should I watch?"

The AI recommended Whiplash (which I'd seen), but also pointed me to Station Eleven, Halt and Catch Fire, and The Knick. Three shows I'd never heard of. All of them were exactly right.

Book Recommendations

Books are where AI recommendations really shine, because book discovery is genuinely broken. Goodreads hasn't been updated in a meaningful way since Amazon bought it. BookTok is fine but trends toward the same 20 books. And asking friends only works if your friends read the same genres you do.

Try this:

"I just read [book] by [author] and it's one of my favorites. I loved [specific thing]. I generally like [genres or styles]. I don't like [dealbreakers]. Recommend 5 books I probably haven't heard of."

That last line is important. Without it, you'll get the obvious picks you've already seen on every "if you liked X" list. Telling the AI to dig deeper actually works.

You can also use AI for something no algorithm does well: cross-media recommendations.

"I loved the movie Arrival. What books would give me a similar feeling?"

Or: "I loved the album 'In Rainbows' by Radiohead. What novels have a similar emotional texture?"

This sounds weird. It works absurdly well. AI understands vibes across formats in a way that separate platform algorithms never will.

Music Discovery

Spotify's algorithm has gotten lazy. If you listen to one genre for a week, it pigeonholes you there for months. And Discover Weekly goes through long stretches of being completely useless.

AI is better for two reasons. First, you can describe what you want in natural language. Second, it can pull from its knowledge of music across decades and genres instead of optimizing for what's trending on the platform right now.

Try this:

"I like [3-5 artists]. What I specifically like about them is [be descriptive]. I want to find artists I've never heard of who have a similar feel. Not mainstream. Give me 10 recommendations with one standout song from each."

Or for playlists:

"I'm hosting a dinner party for 6 people, ages 25-35. The vibe is relaxed but not sleepy. Nobody wants to hear Top 40 but I don't want anything too weird either. Build me a 20-song playlist."

You can copy these song lists directly into Spotify or Apple Music and build the playlist in two minutes.

The Power Move: Building Your Taste Profile

Here's something most people don't think to do. Have a conversation with AI about your taste. Not a single prompt. A conversation.

Start with: "I want to build a taste profile with you. I'm going to tell you about movies, books, shows, and music I love, and I want you to find the patterns in what I like."

Then just list stuff. Your top 10 movies. Five books that changed how you think. Albums you never skip. Shows you've rewatched.

After you've shared enough, ask: "Based on everything I've told you, what patterns do you see in my taste? What am I actually drawn to?"

The answer will probably surprise you. I did this and realized that almost everything I love involves a competent person operating under intense pressure. I'd never noticed that pattern. Now when I look for something new, I know what I'm actually looking for.

And once the AI has that profile, every recommendation it gives you after that is significantly better. It knows your blind spots and your sweet spots.

Which AI to Use

ChatGPT and Claude are both great for this. Gemini is decent but tends to give safer, more obvious picks. If you want surprising, slightly obscure recommendations, Claude edges ahead. If you want fast, reliable picks with good summaries, ChatGPT is solid.

All of them are free to use for this. You don't need a paid plan.

Stop Scrolling

The average American spends about 18 minutes deciding what to watch every time they sit down. That's over 100 hours a year just scrolling through menus. That's genuinely insane.

Next time you can't decide what to watch, read, or listen to, spend 30 seconds talking to AI instead of 30 minutes scrolling through thumbnails. You'll get better picks, discover things you never would have found, and actually enjoy your downtime instead of wasting half of it deciding how to spend it.

If you want more tips on using AI for everyday stuff like this, check out readlaboratories.com/learn or shoot me an email at jake@readlaboratories.com. Always happy to help.