About

Hi, I’m Algon-33 I’m interested in understanding understanding, the foundations of reality and origami. Put a less obnoxious way, I'd like to know how our minds work, how the minds we make work, and how all minds should work. So think computationally bounded Bayesianism, funcitonal neuroscience/figuring out deep learning and logic/decision theory/comptability. That ties into what we can/can't know, which is why I really want to dig into complexity theory. Most all physics is beutiful, but I've had a thing for quantum foundations, chaotic systems and thermodynamics for a while.

I’m currently learning about how to implement a transformer, algorithms to use large language models as writing aids and what to expect out of GPT-4. Also topology: need to prove Brouwer's fixed point theorem. I think I can do it by using ever finer meshes and trying to find open/closed sets whose images overlap with themselves.

I’m looking to collaborate on any ML experiments people find interesting, especially related to interpetability, SGD priors, training path dependancy and grokking.

Or if you want me to do some research for you, I'm game. For $50/hour, you can get me working on a literature review of anything you like. My most recent work was a non-comprehensive [ref]If it were comprehensive, I'd have been working for a year[\ref] survey of ambitious methods of education.

Reach me via email at algon33.ar@gmail.com