Writing
Notes and essays
Most of my writing is on Mechanistic Mind(marked ↗). Short lab notes about projects here live on this site.
- Company Metabolism ↗Smaller animals, like mice, shrews, bats, and sparrows, have faster metabolisms than larger animals like humans, seals, or elephants.
- I'm Picking up their Style ↗Lessons from learning the game of Go
- World Models on the Pareto Frontier ↗Recognize trade-offs through graphs!
- Is Human Memory like RAG? ↗Surprisingly, yes! We can use the differences to forecast what's going to happen with memory in agentic LLMs.
- Links from April ↗Interesting content from the past month
- Having Deep Research Models Review Themselves ↗They get ⭐⭐⭐⭐ out of five!
- Links from March ↗Interesting Ideas and Places from the Internet and from the Past
- Many Common Problems are NP-Hard ↗And Why that Matters for AI
- Claude Code vs Aider ↗Two command line coding assistants: which one is better?
- Using the Terminal — Bash Aliases ↗Some practical advice about using and getting used to the Linux terminal
- Humans are Overparameterized ↗Comparing neurons with Chinchilla scaling laws
- Where I Donated in 2024 ↗In 2024, I split my donations between two grant-making organizations: GiveWell (GW) and Charity Entrepreneurship. Both organizations work in global poverty, but their approaches are quite different, so I split my giving to manage risk.
- Listening to your Conversations and Illustrating them with Generative AI ↗Making a digital picture frame filled with weird AI art
- Numbers without a Zero — Representation in the Brain ↗There is a mystery at the heart of computational neuroscience.
- Diversity of Mechanism in the 👁️👁️ ↗A theme of this blog is that intelligent behavior of humans, animals, and AI systems is made up of mechanisms that we can understand individually.
- Levels of Analysis in Machine Learning ↗There's a concept from cognitive neuroscience of multiple levels of analysis of intelligence. I think this is something that ought to be better known in the world of machine learning, because I find it very helpful for organizing my thoughts.
- Review: Children of Time and the abstract purpose of grammar ↗The book Children of Time features a race of spiders that have developed their own language. What is their language like? How is it structured? And what's the deal with spider webs?
- Hyperbolic Discounting and Pascal's Mugging ↗Google tells you that it is “a psychological bias where people to prioritize immediate rewards and satisfaction over future rewards”. In fact, this is only half right. Hyperbolic discounting also causes people to overvalue distant rewards in addition
- Creating Magic the Gathering cards with generative AI ↗Can generative AI can be used to create novel Magic the Gathering cards? Yes! This is both an article about designing for a card game and about how to work with generative AI for a project like this.
- Contra Alexander on the Bitter Lesson and IQ ↗Scott Alexander argues in favor of treating intelligence as a coherent and somewhat monolithic concept. This is a post about the perils of measurement. I argue that intelligence is multifaceted and it's a mistake to think of it as a linear scale.