A simple model of math skill

Math definitions are easier to understand than theorems, which are easier than proofs. But generating definitions is harder than generating theorems, which are harder than generating proofs.

What agent structure is and why it matters

A 10 minute video talk in which I describe the meaning and motivation behind one of my research projects.

Towards a formalization of the agent structure problem

A semi-technical description of an agent foundations research project that I led in early 2024.

My research agenda in agent foundations

Dealing with infinite entropy

In this second post of the entropy sequence, I show how we can meaningfully understand the entropy of a state even when the number of states is infinite.

My first year in AI alignment

Consider using reversible automata for alignment research

It's popular to use Conway's Game of Life as a model system, but it fails to capture some profound properties of the real world.

"Normal" is the equilibrium state of past optimization processes

Introduction to abstract entropy

The first of a sequence of posts in which I write the best explanation of entropy that I can.

I'm trying out "asteroid mindset"

In which I reflect on how I'm relating to the increasing existential risk from AI.

Giving calibrated time estimates can have social costs

How I became a person who wakes up early

One Year of Pomodoros

Orders of Magnitude

A speech for the 2020 Bay area secular solstice.

The Airstones of Westmining

A little vignette inspired by my partner’s whimsical writing style.

Worship

A future mind decides to experience a piece of humanity's past. [Short story]

Mysterious Light Sources

Light pollution shows us more than just cities.

The International Space Station and an Attempt at Puzzle Hunt Creation

For my 30th birthday party I threw a puzzle hunt with Catherio. I take you through the puzzle so you can solve it yourself, and give instructions for reproducing it for others.

The International Space Station and an Attempt at Gingerbread Modelling

Alone over christmas, I make a gingerbread house, where by “house” I mean the ISS.

The Wound in the World

The International Space Station and an Attempt at Agile Museum Development

Have you ever had an idea so compelling that you didn't know if you would be able to stop yourself from trying to make it real? That happened to me on May 23rd, 2018. [This project didn't get off the ground by the start of the COVID pandemic, and I've now moved on to AI alignment research.]

Just Curious

Reflections on human nature on the 50th anniversary of the Apollo 11 moon landing.

Monument to the Soul

My reaction to the near-loss of Notre Dame de Paris.

The Fragility of Progress

A speech for the 2018 Bay area secular solstice.

Where No Spacecraft Has Gone Before

Introducing the Parker Solar Probe, going closer to the sun than any previous human artifact.

Getting Closer

For my second rocket launch viewing, I bring a telescope to find out what more I can see.

Burning

An ode to that spark inside humanity.

The Maiden Launch of the Falcon Heavy

This was more than a rocket launch. It was a performance, and a shared experience of human achievement.

Call to Arts

In which I urge the rationalists to make more art.

‘But’ Considered Harmful

The Boston CFAR alumni workshop

Simon & Garfunkel and the Natural Order

Weaved throughout their art, I sense a theme of death.

The Playground and the Gameboard

I attempt to give a sense of my subconscious worldview, and how it conflicts with the real world.

The Human Condition

A speech for the 2014 Bay area secular solstice.

One Thing Left to Die

[Poem]

A Comparison of Decision Algorithms on Newcomblike Problems

An Intuitive Explanation of Solomonoff Induction

Solomonoff induction is a mathematical formalization of induction, i.e. predicting the future. The scope of its success is unprecedented and underappreciated. Co-written with Luke Muehlhauser, this piece explains the concept to a general audience.