Hey everyone,

I’ve been experimenting with some AI-based image processing and decided to turn it into a small browser game.

The idea is simple:

- You’re shown an obfuscated movie frame or cover - You try to guess the movie - The image gradually becomes clearer as you use the hints

I built it using Svelte and deployed it on Vercel. This was mainly a learning project to experiment with AI image transformations in a fun way.

I’d really appreciate feedback on: Difficulty balance, UX, Performance or even whether it feels fun or repetitive.

Thanks!

Haha, that was fun. https://i.imgur.com/NBFytY3.png (it has spoilers of which shots the game is using)

I don't have many suggestions besides these:

1. Let the user choose between short, medium and long runs, say 10, 20 and 30 movies.

2. To make this more challenging, ensure the selection includes either exclusively black & white movies, exclusively color movies or a balanced mix of both. 12 Angry Men was easy to identify as it was the only one as such.

3. Perhaps create another ranking for "speedruns" (and differently sized runs).

4. Given that account creation can be a huge barrier for some users, state upfront that you are only requiring username and password.

Also, for some reason I couldn't copy and paste these suggestions in the textarea.