About
Hi! I'm Sudarsh. I'm an incoming freshman at Caltech, where I'm likely going to major in Computational Neuroscience, Physics, Electrical Engineering, or Applied Math.
You can reach me at kvsudarsh786 at gmail.com
You can also read what I've been reading at curius
What I want to do
I want to do cool, hard things. Some of the cool, hard things that I'm interested in right now are
- Neuromorphic computing
- Precise, macro-scale manufacturing
- Better interpretability work on animal brains
- Stateful AI models
Things you can talk to me about!
Cognition
- How do language models update their world model based on context?
- Our brains likely work through making a prediction, and updating based on prediction error. Is this due to sending information back being more expensive, and why?
- Do neurons use backpropogation (current prior: probably not)? If not, what else do they use?
- Why are our brains so unusually sparse? (current hypothesis: energy conservation)
- We can use diffusion models & language models to approximate some amount of human behavior. How far can we push this?
Learning
- How do I learn outside of the framework of university & structured education?
- Where can we find rigor?