You On AI Field Guide · Functional Decision Theory The You On AI Field Guide Home
TxtLowMedHigh
CONCEPT

Functional Decision Theory

The decision theory developed by Nate Soares and Eliezer Yudkowsky that resolves Newcomb-style puzzles by treating a decision as the output of a procedure that may run in more than one place at once, prescribing the choice that does best across all computations of the same function.
Functional decision theory (FDT) is a normative account of how a rational agent ought to choose, introduced by Nate Soares and Eliezer Yudkowsky in a 2017 paper. It competes with causal decision theory (CDT) and evidential decision theory (EDT) by offering what their authors argue is a principled resolution of puzzles about prediction, commitment, and cooperation that CDT and EDT handle poorly. The core move is to reconceive what a decision is. Rather than treating a choice as an isolated physical act with causal downstream effects, FDT treats it as the output of a mathematical function—a decision procedure—that may be instantiated in multiple places simultaneously. When a highly accurate predictor forecasts your choice by simulating your decision procedure, it has, in effect, run that function in advance; your decision therefore logically determines the outcome of every computation of the same function, not merely its physical consequences in the
← Home0%
CONCEPTBook →

Keep reading with YOU ON AI

Unlock the full book, 10,000+ field-guide entries, and a 1000+ thinker library. If you have a book code, register now — it takes a minute.

Register with book code Sign in