Hi everyone. I am making this post to share some work I have been doing for a few months now, as I think it might interest some people on this forum.

My motivation came from trying to find answers to questions like "What is intelligence?" or "What does it really mean to understand?". In looking for answers, I became interested in a field of theoretical computer science called Algorithmic Information Theory. I found it fascinating because it allows for the elegant formalization of everyday concepts that we usually think are impossible to define rigorously. I wanted to explore this more deeply, so I decided to write some lecture notes to synthesize results from different references and also as a personal learning exercise. In my view, AIT provides the best answers we have to those philosophical questions.

Solomonoff’s work and his answer to the problem of induction have been mentioned in various posts and comments on this forum. In my view, Solomonoff’s formalization of Occam’s Razor and his Completeness Theorem represent some of the most beautiful philosophical and mathematical results ever established. Marcus Hutter later extended this paradigm to decision-making, which provided a rigorous foundation for the study of AI. While these concepts are occasionally brought up, I feel they should be a much bigger part of the conversation around AI. This is especially true today, given the current rise of the field.

I am well aware that many people here are experts in these fields and will surely have valid critiques of these notes. I do not claim to be an expert myself. I am simply sharing my process and my synthesis in the hope that they might be useful to someone interested in this approach. You can find the lecture notes, the details of my approach, and the motivations behind this work on the website. I invite you to take a look at it.

www.ait-notes.com

I hope this might interest someone.