Scott Garrabrant

In­tro­duc­tion to Carte­sian Frames

Puz­zle Games

What Would I Do? Self-pre­dic­tion in Sim­ple Algorithms

[Question] Does Agent-like Be­hav­ior Im­ply Agent-like Ar­chi­tec­ture?

In­ten­tional Bucket Errors

Risks from Learned Op­ti­miza­tion: Con­clu­sion and Re­lated Work

De­cep­tive Alignment

The In­ner Align­ment Problem

Con­di­tions for Mesa-Optimization

Risks from Learned Op­ti­miza­tion: Introduction

Yes Re­quires the Pos­si­bil­ity of No

Thoughts on Hu­man Models

Epistemic Tenure

How the MtG Color Wheel Ex­plains AI Safety

[Question] How does Gra­di­ent Des­cent In­ter­act with Good­hart?

For­mal Open Prob­lem in De­ci­sion Theory

The Ubiquitous Con­verse Law­vere Problem

Hyper­real Brouwer

Fixed Point Discussion

Iter­a­tion Fixed Point Exercises