A fun little prototype that uses LLMs + gesture recognition to switch a piece of text between different levels of abstractions: emoji ↔ word ↔ sentence ↔ paragraph ↔ article. Code: https://github.com/chuanenlin/semanticpinching
What if you can pinch your screen to turn an article 📝 into an emoji 💪 and reverse!
I made a fun, simple prototype that uses LLMs + gesture recognition to transform text between different levels of abstractions: emoji ↔ word ↔ sentence ↔ paragraph ↔ article
-
Demo here: https://semanticpinching.vercel.app
Code here: https://github.com/chuanenlin/se...
Twitter thread with examples: https://x.com/chuanenlin/status/...
-
Transforming between modalities could also be interesting: emoji ↔ text ↔ image ↔ video
emoji → text: LLM
text → image: image generation
image → video: video generation
video → image: highlight detection
image → text: image captioning
text → emoji: LLM
Please let me know your thoughts!
Which Frame?