Dual SpaceV*

Working drafts

in other words, incomplete posts that I am not happy with yet

Here is where I keep a few different kinds of working drafts. They should all be somewhat readable, but some may be essentially truncated forms of the final draft, while others may look mostly complete but are in fact supremely deficient in terms of research depth, coherence, and presentation of ideas. I don't necessarily endorse anything you might find here! Tread with great caution, and don't take anything you read here too seriously.

Each page has a stable URL and records when it was first published and last substantially revised. Substantial revisions also appear in the separate working-drafts RSS feed. The main feed remains finished work only.

Working draft

What do we want from utopia?

A childhood wish to do whatever I wanted, and the question of what I actually want from utopia.

First published Last substantially revised #ai #work #philosophy

Working draft

Self-fulfilling info hazards and AI safety

How AI safety rhetoric could feed back into model training and help produce the behavior it warns against.

First published Last substantially revised #ai