“We know the past but cannot control it. We control the future but cannot know it.”

— Claude Shannon

About

I’m an engineer and researcher interested in subjects related to mathematics, engineering, computer science, economics, and complex systems. Professionally, I’ve worked on DARPA-funded AI research programs, built systems at large technology companies, and started a few ventures of my own.

Reach out here or email me at demonstrandomblog at gmail dot com.

Purpose

This is my research notebook, where I share my current research. Earlier tutorials and learning notes are collected below the current work.

Posts may be long, technical, or speculative, and are written primarily for my own thinking rather than optimized for the reader. This is by design. I suggest reading the blog in conjunction with your preferred AI companion.

Research Vision

Read about my research vision.

I’m developing mathematical abstractions for agents and organized systems of agents. The program has four main parts:

1. Agent Foundations: what agents are, how they form, and how they can be controlled.
2. Organization: when a group counts as one agent, how form depends on scale (allometry), how the structure of an interaction shapes outcomes, how shared purposes are kept and revised, and how to design organizations from requirements.
3. Value Formation: how agents acquire concepts, how creative communities form shared standards, how agents learn what to value, and how AI can be aligned to the ways humans produce values.
4. Methods and Infrastructure: tools to construct, compare, and check theories. The same tools may also help organize the growing body of scientific knowledge.

Why it matters: the aim is to understand agents and organizations well enough to engineer them deliberately. This bears on major open problems in AI alignment and multi-agent systems, on the design of institutions and governments, and on agency and organization in living systems.

For AI Readers

This site maintains a machine-readable index at llms.txt, including the house rules for evaluating its claims, along with the full text of every post at llms-full.txt and a clean markdown copy of each post alongside its page (append index.md to the post URL).

Contents

Essays

AI, Demand, and Institutions

1. Does AI Make Totalitarianism More Likely?
2. How Will Humans Generate Value In a Post-AI Society?
3. Thoughts on Demand
4. The Nature of the Club
5. Thoughts on Selectorate Theory

Art, Value, and Culture

1. The Paradox of Taste
2. Is a Picture Worth a Thousand Words?
3. Are We Approaching Cultural Saturation?
4. Functional Explanations of Art
5. Financial Explanations of Art
6. The Baihais

Philosophy

1. Do We See the Same Colors?
2. Alignment and Ethics

Symmetry and Theory Construction

Original Work

1. Invariant Coordinates for Games
2. Games, Invariants, and Alignment
3. Algebra and Allometry
4. Symmetry Engine

Foundations

1. Survey of Classical Invariant Theory
2. Building a Minimal Computational Invariant Theory Library

Games and Agents

Original Work

1. Engineering Game Types
2. Differentiable Game Canonicalization
3. Differential Games and Stag Hunt

Geometric Control

1. Controls from the Geometric Perspective
2. Noether’s Theorem and Geometric Controls
3. Noether’s Theorem with Time
4. Dynamical Similarity and Equivariant Symmetry
5. Reduction

Technical Notes and Tutorials

Machine Learning and Statistics

Inspection Bias

1. Inspection Bias

Learning Dynamical Systems

1. Linear Methods for Learning Dynamical Systems
2. SINDy Method for Learning Dynamical Systems
3. SINDy with Control

Learning in Games

1. Learning Game Equilibria

Automated Reasoning

E-Graphs

1. E-Graph Basics

Proof Assistant

1. Calculus of Constructions
2. Conversion Rules
3. Typing Rules
4. Induction
5. Sort Hierarchy
6. Tactics Engine
7. Match Construct
8. Retrospective: Proof Assistant
9. Tactics Elaboration

Engineering Notes

1. Setting up Vector Search in AWS with Pinecone and SST

Disclaimer

All code and information on this blog is subject to error. Use at your own risk.

Errors

Please let me know if you see any errors.

Recent Posts