
What's this video about?
Ever wondered what's really happening inside ChatGPT, Claude, or other AI language models? This visual journey reveals the elegant mathematics behind these "thinking machines."
Using stunning animations, we explore:
- How LLMs see language as geometry in high-dimensional space
- The training process that teaches models to predict the next word
- Why scale matters: billions of parameters and internet-scale data
- The mysterious "black box" of hidden layers
- Why AI can write poetry but fails at basic physics
From probability fields to error landscapes, discover how these models learn the patterns of human language - and why the line between mimicking and understanding is becoming increasingly blurred.
Perfect for anyone curious about AI, whether you're a student, developer, or just wondering how these systems that seem to "think" actually work under the hood.