I don’t "clock out." I’m 22. I work all day, every day, whenever inspiration strikes. My brain is constantly spinning up new architectures, new stories, and new code. But at 3:30 PM PST, everything stops. 3:30 PM PST That is "Loved Ones Time." It is the only sacred block in my calendar. And because my brain has been running at 100mph since sunrise, the last thing I have the mental bandwidth for is fighting a user interface. We’ve all been there. You sit down to watch a movie, but instead of relaxing, you spend 45 minutes scrolling. The commercial algorithms aren't trying to help you; they are trying to keep you engaged. They want you to hover, click, and watch auto-playing trailers. They also don't understand my taste. If you look at my library, it looks like three different people use my account. I love Ice Cube’s Friday (chill, funny, classic).
I love Camp Rock (Disney musical nostalgia).
I love Bullet Train (high-speed chaos).
I love Wicked and Pitch Perfect. I love Ice Cube’s Friday (chill, funny, classic). Ice Cube’s Friday Friday I love Camp Rock (Disney musical nostalgia). Camp Rock Camp Rock I love Bullet Train (high-speed chaos). Bullet Train Bullet Train I love Wicked and Pitch Perfect. Wicked Wicked Pitch Perfect Pitch Perfect Try explaining to a standard algorithm why I want to watch Brad Pitt fight an assassin on a train immediately after watching Anna Kendrick sing a cappella. The "Recommended for You" tab breaks. It tries to box me into "Action" or "Musicals," but my vibe isn't a genre—it’s Fun. Fun So, I became the Architect. I decided to build a "Baby Algorithm" to curate my 3:30 PM window. The Architect and The Bricklayer I didn't write this alone, but I didn't hire a developer either. I used AI as my "Bricklayer." I have the vision (the Architect). I know that Friday and Camp Rock actually share a similar "comfort watch" DNA despite being polar opposites. I know I hate horror. I know I need trailers. Friday Camp Rock But I didn’t want to spend my precious creative time reading documentation on API response headers. I told the AI what to build, and it laid the cement. It handled the syntax, the loops, and the error handling so I could focus purely on the logic. The Build: Protecting the Vibe I built a Python script using Streamlit (for the UI) and the TMDb API (for the data). The goal was simple: Zero Decision Fatigue. Streamlit TMDb API Zero Decision Fatigue. 1. The "Chaos" Logic Most tutorials tell you to filter by genre. But as I said, my taste is eclectic. If I filtered by "Comedy," I’d lose Bullet Train. If I filtered by "Action," I’d lose Pitch Perfect. Bullet Train Pitch Perfect So, I designed a "Vibe Matcher." It looks at my "Likes" history—this chaotic mix of musicals, stoner comedies, and action flicks—and finds the mathematical center. Then, it applies a "Hard No" filter to protect my peace. "Vibe Matcher." # The "Protective" Filter
discover_url = (
    f"https://api.themoviedb.org/3/discover/movie?api_key={API_KEY}"
    f"&with_genres={most_common_genre_id}" # Finds the common thread
    f"&without_genres=27,53"  # HARD NO to Horror & Thriller
    f"&sort_by=popularity.desc"
) # The "Protective" Filter
discover_url = (
    f"https://api.themoviedb.org/3/discover/movie?api_key={API_KEY}"
    f"&with_genres={most_common_genre_id}" # Finds the common thread
    f"&without_genres=27,53"  # HARD NO to Horror & Thriller
    f"&sort_by=popularity.desc"
) This snippet ensures that no matter how weird my requests get, I never accidentally end up with a stressful thriller during my relaxation time. 2. The Trailer Requirement I’m visual. A poster isn't enough. I need to see the energy. I had the Bricklayer write a function to fetch YouTube links automatically. Now, the app doesn't just suggest a title; it embeds the trailer. I watch 15 seconds. Does it match the energy of Friday? No? Skip. Does it match the fun of Wicked? Yes? Watch. Friday Wicked Watch. The Result: The SpongeBob Victory The true test came yesterday at 3:30 PM. My girlfriend and I (we’re long-distance, so we sync up on PC) opened the app. I fed it our recent favorites to "teach" it. The algorithm crunched the data. It analyzed the music of Camp Rock, the humor of Friday, and the pacing of Bullet Train. Camp Rock Friday Bullet Train It spat out: "The SpongeBob Movie: Search for SquarePants." "The SpongeBob Movie: Search for SquarePants." It was genius. It’s colorful, it has musical numbers, it’s funny, and it’s completely unserious. It was the perfect bridge between our tastes. Why You Should Be An Architect I don't build this stuff because I'm a "coder." I build it because I refuse to let a commercial algorithm dictate my downtime. When you work on inspiration, your downtime is fuel. You can't waste it fighting a UI. By acting as the Architect and using AI as the Bricklayer, I built a tool that respects my eclectic taste and protects my 3:30 PM boundary. It doesn’t judge me for watching Camp Rock at 22. It just gives me the movie and lets me relax. Camp Rock Free yourself from lame shit. Here is the GitHub: https://github.com/damianwgriggs/Movie-Picker-Algo/tree/main https://github.com/damianwgriggs/Movie-Picker-Algo/tree/main

I Built a "Baby Algorithm" to Cure My Decision Fatigue (And Save My Relationship)

About Author

Comments

TOPICS

THIS ARTICLE WAS FEATURED IN

Related Stories

A Quantum AI Oracle: Transmuting Subatomic Entropy into Digital Scarcity

How AI-Powered Decision Intelligence is Transforming Enterprise Cost Optimization

Ethical AI and Autonomous Vehicles: Championing Moral Principles in the Era of Self-Driving Cars

The Data Revolution: AI Takes the Wheel

On the Interpretability of Part-Prototype Based Classifiers: A Human Centric Analysis

Analyzing HIVE Flaws and Part-Prototype Models: A Comparative Study

A Quantum AI Oracle: Transmuting Subatomic Entropy into Digital Scarcity

How AI-Powered Decision Intelligence is Transforming Enterprise Cost Optimization

Ethical AI and Autonomous Vehicles: Championing Moral Principles in the Era of Self-Driving Cars

The Data Revolution: AI Takes the Wheel

On the Interpretability of Part-Prototype Based Classifiers: A Human Centric Analysis

Analyzing HIVE Flaws and Part-Prototype Models: A Comparative Study

Light-Mode

Classic

Newspaper

Minty

Dark-Mode

Neon Noir

Minty

HN StartUps