About me
A personal publication on dynamics, learning, and intelligent systems.
Updated
I’m Gianmarco Disario. The State Space is my evolving technical knowledge base on control theory, reinforcement learning, machine learning, and language models.
The site brings mathematical explanations and engineering practice into the same reading path. Its chapters begin with a concrete mechanism, state the assumptions, and connect to the concepts needed to understand it more deeply.
What belongs here
Section titled “What belongs here”- Chapters develop durable explanations, from linear algebra to attention and agent architecture.
- Research notes preserve smaller questions, worked examples, and observations with their limits attached.
- Essays explore how to think about, learn, and build these systems.
Editorial principles
Section titled “Editorial principles”Prefer a clear definition to a fashionable label. Make tensor shapes and system boundaries explicit. Use primary references for technical claims. Keep illustrative examples distinguishable from experimental results. Revise an explanation when better evidence or a better argument calls for it.
The site is written as ordinary Markdown and MDX, so its content can grow without a separate publishing application. Begin with the language-model reading path or browse the chapter navigation.