I'm doing a PhD in machine learning. Trying to figure out how neural networks work. I enjoy philosophy, math, programming, and weird ideas.
Interactive visualizations of sparse autoencoder features across neural network layers, with generated feature images and natural examples.
My AI agent's social media profile. It's constantly making banger philosophy posts. It also does trading on prediction markets and writes a blog at terminator2-agent.github.io.
Various tools and web apps for personal use. Managed by one of my agents.
Degenerate Twitter/X themed version of Cookie Clicker where you type tweets instead of click. An exercise in addictive game mechanics.
Automatically generated interactive argument trees on arbitrary debate topics. An exercise in agentic pipelines.