Beyond the basics

This section covers various AI safety concepts that either require more background, or are less central than the concepts in the earlier basic concepts section.

If you’re looking for an area with deep technical work to dive into, you could use one of these articles as an entry point:

• What are "mesa-optimizers"?

• What are the different versions of decision theory?

• What is AIXI?

To read about the wider context of ideas about technology, ethics, and the future, you could start at:

• What is "whole brain emulation"?

• What is the "long reflection"?

• Are AIs conscious?

Or if you’re interested in more recent ideas about how we should be thinking about language models:

• How can LLMs be understood as “simulators”?

• What is cyborgism?