Stuart_Armstrong

Re­ward splin­ter­ing as re­verse of interpretability

What are bi­ases, any­way? Mul­ti­ple type signatures

What does GPT-3 un­der­stand? Sym­bol ground­ing and Chi­nese rooms

Re­ward splin­ter­ing for AI design

Bayesi­anism ver­sus con­ser­vatism ver­sus Goodhart

Un­der­ly­ing model of an im­perfect morphism

An­thropic de­ci­sion the­ory for self-lo­cat­ing beliefs

Gen­er­al­ised mod­els: im­perfect mor­phisms and in­for­ma­tional entropy

Prac­ti­cal an­throp­ics summary

An­throp­ics and Fermi: grabby, visi­ble, zoo-keep­ing, and early aliens

The SIA pop­u­la­tion up­date can be sur­pris­ingly small

An­throp­ics in in­finite universes

Non-poi­sonous cake: an­thropic up­dates are normal

The re­verse Good­hart problem

Danger­ous op­ti­mi­sa­tion in­cludes var­i­ance minimisation

The un­der­ly­ing model of a morphism

SIA is ba­si­cally just Bayesian up­dat­ing on existence

The blue-min­imis­ing robot and model splintering

Hu­man pri­ors, fea­tures and mod­els, lan­guages, and Sol­monoff induction

An­throp­ics: differ­ent prob­a­bil­ities, differ­ent questions